feat(executor): refine payload handling by integrating original request context

Updated `applyPayloadConfig` to `applyPayloadConfigWithRoot` across payload translation logic, enabling validation against the original request payload when available. Added support for improved model normalization and translation consistency.

.dockerignore

@@ -13,8 +13,6 @@ Dockerfile
 docs/*
 README.md
 README_CN.md
-MANAGEMENT_API.md
-MANAGEMENT_API_CN.md
 LICENSE
 
 # Runtime data folders (should be mounted as volumes)
@@ -25,10 +23,14 @@ config.yaml
 
 # Development/editor
 bin/*
-.claude/*
 .vscode/*
+.claude/*
+.codex/*
 .gemini/*
 .serena/*
 .agent/*
+.agents/*
+.opencode/*
 .bmad/*
 _bmad/*
+_bmad-output/*

.github/ISSUE_TEMPLATE/bug_report.md

@@ -7,6 +7,13 @@ assignees: ''
 
 ---
 
+**Is it a request payload issue?**
+[  ] Yes, this is a request payload issue. I am using a client/cURL to send a request payload, but I received an unexpected error.
+[  ] No, it's another issue.
+
+**If it's a request payload issue, you MUST know**
+Our team doesn't have any GODs or ORACLEs or MIND READERs. Please make sure to attach the request log or curl payload.
+
 **Describe the bug**
 A clear and concise description of what the bug is.
 

.gitignore

@@ -12,11 +12,15 @@ bin/*
 logs/*
 conv/*
 temp/*
+refs/*
+
+# Storage backends
 pgstore/*
 gitstore/*
 objectstore/*
+
+# Static assets
 static/*
-refs/*
 
 # Authentication data
 auths/*
@@ -30,12 +34,17 @@ GEMINI.md
 
 # Tooling metadata
 .vscode/*
+.codex/*
 .claude/*
 .gemini/*
 .serena/*
 .agent/*
+.agents/*
+.agents/*
+.opencode/*
 .bmad/*
 _bmad/*
+_bmad-output/*
 .mcp/cache/
 
 # macOS

cmd/server/main.go

@@ -434,7 +434,7 @@ func main() {
 	usage.SetStatisticsEnabled(cfg.UsageStatisticsEnabled)
 	coreauth.SetQuotaCooldownDisabled(cfg.DisableCooling)
 
-	if err = logging.ConfigureLogOutput(cfg.LoggingToFile, cfg.LogsMaxTotalSizeMB); err != nil {
+	if err = logging.ConfigureLogOutput(cfg); err != nil {
 		log.Errorf("failed to configure log output: %v", err)
 		return
 	}

config.example.yaml

@@ -35,10 +35,14 @@ auth-dir: "~/.cli-proxy-api"
 api-keys:
   - "your-api-key-1"
   - "your-api-key-2"
+  - "your-api-key-3"
 
 # Enable debug logging
 debug: false
 
+# When true, disable high-overhead HTTP middleware features to reduce per-request memory usage under high concurrency.
+commercial-mode: false
+
 # Open OAuth URLs in incognito/private browser mode.
 # Useful when you want to login with a different account without logging out from your current session.
 # Default: false (but Kiro auth defaults to true for multi-account support)
@@ -71,9 +75,18 @@ quota-exceeded:
   switch-project: true # Whether to automatically switch to another project when a quota is exceeded
   switch-preview-model: true # Whether to automatically switch to a preview model when a quota is exceeded
 
+# Routing strategy for selecting credentials when multiple match.
+routing:
+  strategy: "round-robin" # round-robin (default), fill-first
+
 # When true, enable authentication for the WebSocket API (/v1/ws).
 ws-auth: false
 
+# Streaming behavior (SSE keep-alives + safe bootstrap retries).
+# streaming:
+#   keepalive-seconds: 15   # Default: 0 (disabled). <= 0 disables keep-alives.
+#   bootstrap-retries: 1    # Default: 0 (disabled). Retries before first byte is sent.
+
 # Gemini API keys
 # gemini-api-key:
 #   - api-key: "AIzaSy...01"
@@ -82,6 +95,9 @@ ws-auth: false
 #     headers:
 #       X-Custom-Header: "custom-value"
 #     proxy-url: "socks5://proxy.example.com:1080"
+#     models:
+#       - name: "gemini-2.5-flash" # upstream model name
+#         alias: "gemini-flash"    # client alias mapped to the upstream model
 #     excluded-models:
 #       - "gemini-2.5-pro"     # exclude specific models from this provider (exact match)
 #       - "gemini-2.5-*"       # wildcard matching prefix (e.g. gemini-2.5-flash, gemini-2.5-pro)
@@ -97,6 +113,9 @@ ws-auth: false
 #     headers:
 #       X-Custom-Header: "custom-value"
 #     proxy-url: "socks5://proxy.example.com:1080" # optional: per-key proxy override
+#     models:
+#       - name: "gpt-5-codex"   # upstream model name
+#         alias: "codex-latest" # client alias mapped to the upstream model
 #     excluded-models:
 #       - "gpt-5.1"         # exclude specific models (exact match)
 #       - "gpt-5-*"         # wildcard matching prefix (e.g. gpt-5-medium, gpt-5-codex)
@@ -114,7 +133,7 @@ ws-auth: false
 #     proxy-url: "socks5://proxy.example.com:1080" # optional: per-key proxy override
 #     models:
 #       - name: "claude-3-5-sonnet-20241022" # upstream model name
-#         alias: "claude-sonnet-latest" # client alias mapped to the upstream model
+#         alias: "claude-sonnet-latest"      # client alias mapped to the upstream model
 #     excluded-models:
 #       - "claude-opus-4-5-20251101" # exclude specific models (exact match)
 #       - "claude-3-*"               # wildcard matching prefix (e.g. claude-3-7-sonnet-20250219)
@@ -155,9 +174,9 @@ ws-auth: false
 #     headers:
 #       X-Custom-Header: "custom-value"
 #     models:                                     # optional: map aliases to upstream model names
-#       - name: "gemini-2.0-flash"                # upstream model name
+#       - name: "gemini-2.5-flash"                # upstream model name
 #         alias: "vertex-flash"                   # client-visible alias
-#       - name: "gemini-1.5-pro"
+#       - name: "gemini-2.5-pro"
 #         alias: "vertex-pro"
 
 # Amp Integration
@@ -166,6 +185,18 @@ ws-auth: false
 #   upstream-url: "https://ampcode.com"
 #   # Optional: Override API key for Amp upstream (otherwise uses env or file)
 #   upstream-api-key: ""
+#   # Per-client upstream API key mapping
+#   # Maps client API keys (from top-level api-keys) to different Amp upstream API keys.
+#   # Useful when different clients need to use different Amp accounts/quotas.
+#   # If a client key isn't mapped, falls back to upstream-api-key (default behavior).
+#   upstream-api-keys:
+#     - upstream-api-key: "amp_key_for_team_a"    # Upstream key to use for these clients
+#       api-keys:                                 # Client keys that use this upstream key
+#         - "your-api-key-1"
+#         - "your-api-key-2"
+#     - upstream-api-key: "amp_key_for_team_b"
+#       api-keys:
+#         - "your-api-key-3"
 #   # Restrict Amp management routes (/api/auth, /api/user, etc.) to localhost only (default: false)
 #   restrict-management-to-localhost: false
 #   # Force model mappings to run before checking local API keys (default: false)
@@ -175,12 +206,42 @@ ws-auth: false
 #   # Useful when Amp CLI requests models you don't have access to (e.g., Claude Opus 4.5)
 #   # but you have a similar model available (e.g., Claude Sonnet 4).
 #   model-mappings:
-#     - from: "claude-opus-4.5"       # Model requested by Amp CLI
-#       to: "claude-sonnet-4"         # Route to this available model instead
-#     - from: "gpt-5"
-#       to: "gemini-2.5-pro"
-#     - from: "claude-3-opus-20240229"
-#       to: "claude-3-5-sonnet-20241022"
+#     - from: "claude-opus-4-5-20251101"          # Model requested by Amp CLI
+#       to: "gemini-claude-opus-4-5-thinking"     # Route to this available model instead
+#     - from: "claude-sonnet-4-5-20250929"
+#       to: "gemini-claude-sonnet-4-5-thinking"
+#     - from: "claude-haiku-4-5-20251001"
+#       to: "gemini-2.5-flash"
+
+# Global OAuth model name mappings (per channel)
+# These mappings rename model IDs for both model listing and request routing.
+# Supported channels: gemini-cli, vertex, aistudio, antigravity, claude, codex, qwen, iflow.
+# NOTE: Mappings do not apply to gemini-api-key, codex-api-key, claude-api-key, openai-compatibility, vertex-api-key, or ampcode.
+# oauth-model-mappings:
+#   gemini-cli:
+#     - name: "gemini-2.5-pro"          # original model name under this channel
+#       alias: "g2.5p"                  # client-visible alias
+#   vertex:
+#     - name: "gemini-2.5-pro"
+#       alias: "g2.5p"
+#   aistudio:
+#     - name: "gemini-2.5-pro"
+#       alias: "g2.5p"
+#   antigravity:
+#     - name: "gemini-3-pro-preview"
+#       alias: "g3p"
+#   claude:
+#     - name: "claude-sonnet-4-5-20250929"
+#       alias: "cs4.5"
+#   codex:
+#     - name: "gpt-5"
+#       alias: "g5"
+#   qwen:
+#     - name: "qwen3-coder-plus"
+#       alias: "qwen-plus"
+#   iflow:
+#     - name: "glm-4.7"
+#       alias: "glm-god"
 
 # OAuth provider excluded models
 # oauth-excluded-models:

internal/api/handlers/management/api_tools.go

@@ -0,0 +1,538 @@
+package management
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"io"
+	"net"
+	"net/http"
+	"net/url"
+	"strings"
+	"time"
+
+	"github.com/gin-gonic/gin"
+	"github.com/router-for-me/CLIProxyAPI/v6/internal/runtime/geminicli"
+	coreauth "github.com/router-for-me/CLIProxyAPI/v6/sdk/cliproxy/auth"
+	log "github.com/sirupsen/logrus"
+	"golang.org/x/net/proxy"
+	"golang.org/x/oauth2"
+	"golang.org/x/oauth2/google"
+)
+
+const defaultAPICallTimeout = 60 * time.Second
+
+const (
+	geminiOAuthClientID     = "681255809395-oo8ft2oprdrnp9e3aqf6av3hmdib135j.apps.googleusercontent.com"
+	geminiOAuthClientSecret = "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"
+)
+
+var geminiOAuthScopes = []string{
+	"https://www.googleapis.com/auth/cloud-platform",
+	"https://www.googleapis.com/auth/userinfo.email",
+	"https://www.googleapis.com/auth/userinfo.profile",
+}
+
+type apiCallRequest struct {
+	AuthIndexSnake  *string           `json:"auth_index"`
+	AuthIndexCamel  *string           `json:"authIndex"`
+	AuthIndexPascal *string           `json:"AuthIndex"`
+	Method          string            `json:"method"`
+	URL             string            `json:"url"`
+	Header          map[string]string `json:"header"`
+	Data            string            `json:"data"`
+}
+
+type apiCallResponse struct {
+	StatusCode int                 `json:"status_code"`
+	Header     map[string][]string `json:"header"`
+	Body       string              `json:"body"`
+}
+
+// APICall makes a generic HTTP request on behalf of the management API caller.
+// It is protected by the management middleware.
+//
+// Endpoint:
+//
+//	POST /v0/management/api-call
+//
+// Authentication:
+//
+//	Same as other management APIs (requires a management key and remote-management rules).
+//	You can provide the key via:
+//	- Authorization: Bearer <key>
+//	- X-Management-Key: <key>
+//
+// Request JSON:
+//   - auth_index / authIndex / AuthIndex (optional):
+//     The credential "auth_index" from GET /v0/management/auth-files (or other endpoints returning it).
+//     If omitted or not found, credential-specific proxy/token substitution is skipped.
+//   - method (required): HTTP method, e.g. GET, POST, PUT, PATCH, DELETE.
+//   - url (required): Absolute URL including scheme and host, e.g. "https://api.example.com/v1/ping".
+//   - header (optional): Request headers map.
+//     Supports magic variable "$TOKEN$" which is replaced using the selected credential:
+//     1) metadata.access_token
+//     2) attributes.api_key
+//     3) metadata.token / metadata.id_token / metadata.cookie
+//     Example: {"Authorization":"Bearer $TOKEN$"}.
+//     Note: if you need to override the HTTP Host header, set header["Host"].
+//   - data (optional): Raw request body as string (useful for POST/PUT/PATCH).
+//
+// Proxy selection (highest priority first):
+//  1. Selected credential proxy_url
+//  2. Global config proxy-url
+//  3. Direct connect (environment proxies are not used)
+//
+// Response JSON (returned with HTTP 200 when the APICall itself succeeds):
+//   - status_code: Upstream HTTP status code.
+//   - header: Upstream response headers.
+//   - body: Upstream response body as string.
+//
+// Example:
+//
+//	curl -sS -X POST "http://127.0.0.1:8317/v0/management/api-call" \
+//	  -H "Authorization: Bearer <MANAGEMENT_KEY>" \
+//	  -H "Content-Type: application/json" \
+//	  -d '{"auth_index":"<AUTH_INDEX>","method":"GET","url":"https://api.example.com/v1/ping","header":{"Authorization":"Bearer $TOKEN$"}}'
+//
+//	curl -sS -X POST "http://127.0.0.1:8317/v0/management/api-call" \
+//	  -H "Authorization: Bearer 831227" \
+//	  -H "Content-Type: application/json" \
+//	  -d '{"auth_index":"<AUTH_INDEX>","method":"POST","url":"https://api.example.com/v1/fetchAvailableModels","header":{"Authorization":"Bearer $TOKEN$","Content-Type":"application/json","User-Agent":"cliproxyapi"},"data":"{}"}'
+func (h *Handler) APICall(c *gin.Context) {
+	var body apiCallRequest
+	if errBindJSON := c.ShouldBindJSON(&body); errBindJSON != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "invalid body"})
+		return
+	}
+
+	method := strings.ToUpper(strings.TrimSpace(body.Method))
+	if method == "" {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "missing method"})
+		return
+	}
+
+	urlStr := strings.TrimSpace(body.URL)
+	if urlStr == "" {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "missing url"})
+		return
+	}
+	parsedURL, errParseURL := url.Parse(urlStr)
+	if errParseURL != nil || parsedURL.Scheme == "" || parsedURL.Host == "" {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "invalid url"})
+		return
+	}
+
+	authIndex := firstNonEmptyString(body.AuthIndexSnake, body.AuthIndexCamel, body.AuthIndexPascal)
+	auth := h.authByIndex(authIndex)
+
+	reqHeaders := body.Header
+	if reqHeaders == nil {
+		reqHeaders = map[string]string{}
+	}
+
+	var hostOverride string
+	var token string
+	var tokenResolved bool
+	var tokenErr error
+	for key, value := range reqHeaders {
+		if !strings.Contains(value, "$TOKEN$") {
+			continue
+		}
+		if !tokenResolved {
+			token, tokenErr = h.resolveTokenForAuth(c.Request.Context(), auth)
+			tokenResolved = true
+		}
+		if auth != nil && token == "" {
+			if tokenErr != nil {
+				c.JSON(http.StatusBadRequest, gin.H{"error": "auth token refresh failed"})
+				return
+			}
+			c.JSON(http.StatusBadRequest, gin.H{"error": "auth token not found"})
+			return
+		}
+		if token == "" {
+			continue
+		}
+		reqHeaders[key] = strings.ReplaceAll(value, "$TOKEN$", token)
+	}
+
+	var requestBody io.Reader
+	if body.Data != "" {
+		requestBody = strings.NewReader(body.Data)
+	}
+
+	req, errNewRequest := http.NewRequestWithContext(c.Request.Context(), method, urlStr, requestBody)
+	if errNewRequest != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "failed to build request"})
+		return
+	}
+
+	for key, value := range reqHeaders {
+		if strings.EqualFold(key, "host") {
+			hostOverride = strings.TrimSpace(value)
+			continue
+		}
+		req.Header.Set(key, value)
+	}
+	if hostOverride != "" {
+		req.Host = hostOverride
+	}
+
+	httpClient := &http.Client{
+		Timeout: defaultAPICallTimeout,
+	}
+	httpClient.Transport = h.apiCallTransport(auth)
+
+	resp, errDo := httpClient.Do(req)
+	if errDo != nil {
+		log.WithError(errDo).Debug("management APICall request failed")
+		c.JSON(http.StatusBadGateway, gin.H{"error": "request failed"})
+		return
+	}
+	defer func() {
+		if errClose := resp.Body.Close(); errClose != nil {
+			log.Errorf("response body close error: %v", errClose)
+		}
+	}()
+
+	respBody, errReadAll := io.ReadAll(resp.Body)
+	if errReadAll != nil {
+		c.JSON(http.StatusBadGateway, gin.H{"error": "failed to read response"})
+		return
+	}
+
+	c.JSON(http.StatusOK, apiCallResponse{
+		StatusCode: resp.StatusCode,
+		Header:     resp.Header,
+		Body:       string(respBody),
+	})
+}
+
+func firstNonEmptyString(values ...*string) string {
+	for _, v := range values {
+		if v == nil {
+			continue
+		}
+		if out := strings.TrimSpace(*v); out != "" {
+			return out
+		}
+	}
+	return ""
+}
+
+func tokenValueForAuth(auth *coreauth.Auth) string {
+	if auth == nil {
+		return ""
+	}
+	if v := tokenValueFromMetadata(auth.Metadata); v != "" {
+		return v
+	}
+	if auth.Attributes != nil {
+		if v := strings.TrimSpace(auth.Attributes["api_key"]); v != "" {
+			return v
+		}
+	}
+	if shared := geminicli.ResolveSharedCredential(auth.Runtime); shared != nil {
+		if v := tokenValueFromMetadata(shared.MetadataSnapshot()); v != "" {
+			return v
+		}
+	}
+	return ""
+}
+
+func (h *Handler) resolveTokenForAuth(ctx context.Context, auth *coreauth.Auth) (string, error) {
+	if auth == nil {
+		return "", nil
+	}
+
+	provider := strings.ToLower(strings.TrimSpace(auth.Provider))
+	if provider == "gemini-cli" {
+		token, errToken := h.refreshGeminiOAuthAccessToken(ctx, auth)
+		return token, errToken
+	}
+
+	return tokenValueForAuth(auth), nil
+}
+
+func (h *Handler) refreshGeminiOAuthAccessToken(ctx context.Context, auth *coreauth.Auth) (string, error) {
+	if ctx == nil {
+		ctx = context.Background()
+	}
+	if auth == nil {
+		return "", nil
+	}
+
+	metadata, updater := geminiOAuthMetadata(auth)
+	if len(metadata) == 0 {
+		return "", fmt.Errorf("gemini oauth metadata missing")
+	}
+
+	base := make(map[string]any)
+	if tokenRaw, ok := metadata["token"].(map[string]any); ok && tokenRaw != nil {
+		base = cloneMap(tokenRaw)
+	}
+
+	var token oauth2.Token
+	if len(base) > 0 {
+		if raw, errMarshal := json.Marshal(base); errMarshal == nil {
+			_ = json.Unmarshal(raw, &token)
+		}
+	}
+
+	if token.AccessToken == "" {
+		token.AccessToken = stringValue(metadata, "access_token")
+	}
+	if token.RefreshToken == "" {
+		token.RefreshToken = stringValue(metadata, "refresh_token")
+	}
+	if token.TokenType == "" {
+		token.TokenType = stringValue(metadata, "token_type")
+	}
+	if token.Expiry.IsZero() {
+		if expiry := stringValue(metadata, "expiry"); expiry != "" {
+			if ts, errParseTime := time.Parse(time.RFC3339, expiry); errParseTime == nil {
+				token.Expiry = ts
+			}
+		}
+	}
+
+	conf := &oauth2.Config{
+		ClientID:     geminiOAuthClientID,
+		ClientSecret: geminiOAuthClientSecret,
+		Scopes:       geminiOAuthScopes,
+		Endpoint:     google.Endpoint,
+	}
+
+	ctxToken := ctx
+	httpClient := &http.Client{
+		Timeout:   defaultAPICallTimeout,
+		Transport: h.apiCallTransport(auth),
+	}
+	ctxToken = context.WithValue(ctxToken, oauth2.HTTPClient, httpClient)
+
+	src := conf.TokenSource(ctxToken, &token)
+	currentToken, errToken := src.Token()
+	if errToken != nil {
+		return "", errToken
+	}
+
+	merged := buildOAuthTokenMap(base, currentToken)
+	fields := buildOAuthTokenFields(currentToken, merged)
+	if updater != nil {
+		updater(fields)
+	}
+	return strings.TrimSpace(currentToken.AccessToken), nil
+}
+
+func geminiOAuthMetadata(auth *coreauth.Auth) (map[string]any, func(map[string]any)) {
+	if auth == nil {
+		return nil, nil
+	}
+	if shared := geminicli.ResolveSharedCredential(auth.Runtime); shared != nil {
+		snapshot := shared.MetadataSnapshot()
+		return snapshot, func(fields map[string]any) { shared.MergeMetadata(fields) }
+	}
+	return auth.Metadata, func(fields map[string]any) {
+		if auth.Metadata == nil {
+			auth.Metadata = make(map[string]any)
+		}
+		for k, v := range fields {
+			auth.Metadata[k] = v
+		}
+	}
+}
+
+func stringValue(metadata map[string]any, key string) string {
+	if len(metadata) == 0 || key == "" {
+		return ""
+	}
+	if v, ok := metadata[key].(string); ok {
+		return strings.TrimSpace(v)
+	}
+	return ""
+}
+
+func cloneMap(in map[string]any) map[string]any {
+	if len(in) == 0 {
+		return nil
+	}
+	out := make(map[string]any, len(in))
+	for k, v := range in {
+		out[k] = v
+	}
+	return out
+}
+
+func buildOAuthTokenMap(base map[string]any, tok *oauth2.Token) map[string]any {
+	merged := cloneMap(base)
+	if merged == nil {
+		merged = make(map[string]any)
+	}
+	if tok == nil {
+		return merged
+	}
+	if raw, errMarshal := json.Marshal(tok); errMarshal == nil {
+		var tokenMap map[string]any
+		if errUnmarshal := json.Unmarshal(raw, &tokenMap); errUnmarshal == nil {
+			for k, v := range tokenMap {
+				merged[k] = v
+			}
+		}
+	}
+	return merged
+}
+
+func buildOAuthTokenFields(tok *oauth2.Token, merged map[string]any) map[string]any {
+	fields := make(map[string]any, 5)
+	if tok != nil && tok.AccessToken != "" {
+		fields["access_token"] = tok.AccessToken
+	}
+	if tok != nil && tok.TokenType != "" {
+		fields["token_type"] = tok.TokenType
+	}
+	if tok != nil && tok.RefreshToken != "" {
+		fields["refresh_token"] = tok.RefreshToken
+	}
+	if tok != nil && !tok.Expiry.IsZero() {
+		fields["expiry"] = tok.Expiry.Format(time.RFC3339)
+	}
+	if len(merged) > 0 {
+		fields["token"] = cloneMap(merged)
+	}
+	return fields
+}
+
+func tokenValueFromMetadata(metadata map[string]any) string {
+	if len(metadata) == 0 {
+		return ""
+	}
+	if v, ok := metadata["accessToken"].(string); ok && strings.TrimSpace(v) != "" {
+		return strings.TrimSpace(v)
+	}
+	if v, ok := metadata["access_token"].(string); ok && strings.TrimSpace(v) != "" {
+		return strings.TrimSpace(v)
+	}
+	if tokenRaw, ok := metadata["token"]; ok && tokenRaw != nil {
+		switch typed := tokenRaw.(type) {
+		case string:
+			if v := strings.TrimSpace(typed); v != "" {
+				return v
+			}
+		case map[string]any:
+			if v, ok := typed["access_token"].(string); ok && strings.TrimSpace(v) != "" {
+				return strings.TrimSpace(v)
+			}
+			if v, ok := typed["accessToken"].(string); ok && strings.TrimSpace(v) != "" {
+				return strings.TrimSpace(v)
+			}
+		case map[string]string:
+			if v := strings.TrimSpace(typed["access_token"]); v != "" {
+				return v
+			}
+			if v := strings.TrimSpace(typed["accessToken"]); v != "" {
+				return v
+			}
+		}
+	}
+	if v, ok := metadata["token"].(string); ok && strings.TrimSpace(v) != "" {
+		return strings.TrimSpace(v)
+	}
+	if v, ok := metadata["id_token"].(string); ok && strings.TrimSpace(v) != "" {
+		return strings.TrimSpace(v)
+	}
+	if v, ok := metadata["cookie"].(string); ok && strings.TrimSpace(v) != "" {
+		return strings.TrimSpace(v)
+	}
+	return ""
+}
+
+func (h *Handler) authByIndex(authIndex string) *coreauth.Auth {
+	authIndex = strings.TrimSpace(authIndex)
+	if authIndex == "" || h == nil || h.authManager == nil {
+		return nil
+	}
+	auths := h.authManager.List()
+	for _, auth := range auths {
+		if auth == nil {
+			continue
+		}
+		auth.EnsureIndex()
+		if auth.Index == authIndex {
+			return auth
+		}
+	}
+	return nil
+}
+
+func (h *Handler) apiCallTransport(auth *coreauth.Auth) http.RoundTripper {
+	var proxyCandidates []string
+	if auth != nil {
+		if proxyStr := strings.TrimSpace(auth.ProxyURL); proxyStr != "" {
+			proxyCandidates = append(proxyCandidates, proxyStr)
+		}
+	}
+	if h != nil && h.cfg != nil {
+		if proxyStr := strings.TrimSpace(h.cfg.ProxyURL); proxyStr != "" {
+			proxyCandidates = append(proxyCandidates, proxyStr)
+		}
+	}
+
+	for _, proxyStr := range proxyCandidates {
+		if transport := buildProxyTransport(proxyStr); transport != nil {
+			return transport
+		}
+	}
+
+	transport, ok := http.DefaultTransport.(*http.Transport)
+	if !ok || transport == nil {
+		return &http.Transport{Proxy: nil}
+	}
+	clone := transport.Clone()
+	clone.Proxy = nil
+	return clone
+}
+
+func buildProxyTransport(proxyStr string) *http.Transport {
+	proxyStr = strings.TrimSpace(proxyStr)
+	if proxyStr == "" {
+		return nil
+	}
+
+	proxyURL, errParse := url.Parse(proxyStr)
+	if errParse != nil {
+		log.WithError(errParse).Debug("parse proxy URL failed")
+		return nil
+	}
+	if proxyURL.Scheme == "" || proxyURL.Host == "" {
+		log.Debug("proxy URL missing scheme/host")
+		return nil
+	}
+
+	if proxyURL.Scheme == "socks5" {
+		var proxyAuth *proxy.Auth
+		if proxyURL.User != nil {
+			username := proxyURL.User.Username()
+			password, _ := proxyURL.User.Password()
+			proxyAuth = &proxy.Auth{User: username, Password: password}
+		}
+		dialer, errSOCKS5 := proxy.SOCKS5("tcp", proxyURL.Host, proxyAuth, proxy.Direct)
+		if errSOCKS5 != nil {
+			log.WithError(errSOCKS5).Debug("create SOCKS5 dialer failed")
+			return nil
+		}
+		return &http.Transport{
+			Proxy: nil,
+			DialContext: func(ctx context.Context, network, addr string) (net.Conn, error) {
+				return dialer.Dial(network, addr)
+			},
+		}
+	}
+
+	if proxyURL.Scheme == "http" || proxyURL.Scheme == "https" {
+		return &http.Transport{Proxy: http.ProxyURL(proxyURL)}
+	}
+
+	log.Debugf("unsupported proxy scheme: %s", proxyURL.Scheme)
+	return nil
+}

internal/api/handlers/management/auth_files.go

@@ -431,9 +431,46 @@ func (h *Handler) buildAuthFileEntry(auth *coreauth.Auth) gin.H {
 			log.WithError(err).Warnf("failed to stat auth file %s", path)
 		}
 	}
+	if claims := extractCodexIDTokenClaims(auth); claims != nil {
+		entry["id_token"] = claims
+	}
 	return entry
 }
 
+func extractCodexIDTokenClaims(auth *coreauth.Auth) gin.H {
+	if auth == nil || auth.Metadata == nil {
+		return nil
+	}
+	if !strings.EqualFold(strings.TrimSpace(auth.Provider), "codex") {
+		return nil
+	}
+	idTokenRaw, ok := auth.Metadata["id_token"].(string)
+	if !ok {
+		return nil
+	}
+	idToken := strings.TrimSpace(idTokenRaw)
+	if idToken == "" {
+		return nil
+	}
+	claims, err := codex.ParseJWTToken(idToken)
+	if err != nil || claims == nil {
+		return nil
+	}
+
+	result := gin.H{}
+	if v := strings.TrimSpace(claims.CodexAuthInfo.ChatgptAccountID); v != "" {
+		result["chatgpt_account_id"] = v
+	}
+	if v := strings.TrimSpace(claims.CodexAuthInfo.ChatgptPlanType); v != "" {
+		result["plan_type"] = v
+	}
+
+	if len(result) == 0 {
+		return nil
+	}
+	return result
+}
+
 func authEmail(auth *coreauth.Auth) string {
 	if auth == nil {
 		return ""

internal/api/handlers/management/config_lists.go

@@ -597,11 +597,7 @@ func (h *Handler) PutCodexKeys(c *gin.Context) {
 	filtered := make([]config.CodexKey, 0, len(arr))
 	for i := range arr {
 		entry := arr[i]
-		entry.APIKey = strings.TrimSpace(entry.APIKey)
-		entry.BaseURL = strings.TrimSpace(entry.BaseURL)
-		entry.ProxyURL = strings.TrimSpace(entry.ProxyURL)
-		entry.Headers = config.NormalizeHeaders(entry.Headers)
-		entry.ExcludedModels = config.NormalizeExcludedModels(entry.ExcludedModels)
+		normalizeCodexKey(&entry)
 		if entry.BaseURL == "" {
 			continue
 		}
@@ -613,12 +609,13 @@ func (h *Handler) PutCodexKeys(c *gin.Context) {
 }
 func (h *Handler) PatchCodexKey(c *gin.Context) {
 	type codexKeyPatch struct {
-		APIKey         *string            `json:"api-key"`
-		Prefix         *string            `json:"prefix"`
-		BaseURL        *string            `json:"base-url"`
-		ProxyURL       *string            `json:"proxy-url"`
-		Headers        *map[string]string `json:"headers"`
-		ExcludedModels *[]string          `json:"excluded-models"`
+		APIKey         *string              `json:"api-key"`
+		Prefix         *string              `json:"prefix"`
+		BaseURL        *string              `json:"base-url"`
+		ProxyURL       *string              `json:"proxy-url"`
+		Models         *[]config.CodexModel `json:"models"`
+		Headers        *map[string]string   `json:"headers"`
+		ExcludedModels *[]string            `json:"excluded-models"`
 	}
 	var body struct {
 		Index *int           `json:"index"`
@@ -667,12 +664,16 @@ func (h *Handler) PatchCodexKey(c *gin.Context) {
 	if body.Value.ProxyURL != nil {
 		entry.ProxyURL = strings.TrimSpace(*body.Value.ProxyURL)
 	}
+	if body.Value.Models != nil {
+		entry.Models = append([]config.CodexModel(nil), (*body.Value.Models)...)
+	}
 	if body.Value.Headers != nil {
 		entry.Headers = config.NormalizeHeaders(*body.Value.Headers)
 	}
 	if body.Value.ExcludedModels != nil {
 		entry.ExcludedModels = config.NormalizeExcludedModels(*body.Value.ExcludedModels)
 	}
+	normalizeCodexKey(&entry)
 	h.cfg.CodexKey[targetIndex] = entry
 	h.cfg.SanitizeCodexKeys()
 	h.persist(c)
@@ -762,6 +763,32 @@ func normalizeClaudeKey(entry *config.ClaudeKey) {
 	entry.Models = normalized
 }
 
+func normalizeCodexKey(entry *config.CodexKey) {
+	if entry == nil {
+		return
+	}
+	entry.APIKey = strings.TrimSpace(entry.APIKey)
+	entry.Prefix = strings.TrimSpace(entry.Prefix)
+	entry.BaseURL = strings.TrimSpace(entry.BaseURL)
+	entry.ProxyURL = strings.TrimSpace(entry.ProxyURL)
+	entry.Headers = config.NormalizeHeaders(entry.Headers)
+	entry.ExcludedModels = config.NormalizeExcludedModels(entry.ExcludedModels)
+	if len(entry.Models) == 0 {
+		return
+	}
+	normalized := make([]config.CodexModel, 0, len(entry.Models))
+	for i := range entry.Models {
+		model := entry.Models[i]
+		model.Name = strings.TrimSpace(model.Name)
+		model.Alias = strings.TrimSpace(model.Alias)
+		if model.Name == "" && model.Alias == "" {
+			continue
+		}
+		normalized = append(normalized, model)
+	}
+	entry.Models = normalized
+}
+
 // GetAmpCode returns the complete ampcode configuration.
 func (h *Handler) GetAmpCode(c *gin.Context) {
 	if h == nil || h.cfg == nil {
@@ -913,3 +940,151 @@ func (h *Handler) GetAmpForceModelMappings(c *gin.Context) {
 func (h *Handler) PutAmpForceModelMappings(c *gin.Context) {
 	h.updateBoolField(c, func(v bool) { h.cfg.AmpCode.ForceModelMappings = v })
 }
+
+// GetAmpUpstreamAPIKeys returns the ampcode upstream API keys mapping.
+func (h *Handler) GetAmpUpstreamAPIKeys(c *gin.Context) {
+	if h == nil || h.cfg == nil {
+		c.JSON(200, gin.H{"upstream-api-keys": []config.AmpUpstreamAPIKeyEntry{}})
+		return
+	}
+	c.JSON(200, gin.H{"upstream-api-keys": h.cfg.AmpCode.UpstreamAPIKeys})
+}
+
+// PutAmpUpstreamAPIKeys replaces all ampcode upstream API keys mappings.
+func (h *Handler) PutAmpUpstreamAPIKeys(c *gin.Context) {
+	var body struct {
+		Value []config.AmpUpstreamAPIKeyEntry `json:"value"`
+	}
+	if err := c.ShouldBindJSON(&body); err != nil {
+		c.JSON(400, gin.H{"error": "invalid body"})
+		return
+	}
+	// Normalize entries: trim whitespace, filter empty
+	normalized := normalizeAmpUpstreamAPIKeyEntries(body.Value)
+	h.cfg.AmpCode.UpstreamAPIKeys = normalized
+	h.persist(c)
+}
+
+// PatchAmpUpstreamAPIKeys adds or updates upstream API keys entries.
+// Matching is done by upstream-api-key value.
+func (h *Handler) PatchAmpUpstreamAPIKeys(c *gin.Context) {
+	var body struct {
+		Value []config.AmpUpstreamAPIKeyEntry `json:"value"`
+	}
+	if err := c.ShouldBindJSON(&body); err != nil {
+		c.JSON(400, gin.H{"error": "invalid body"})
+		return
+	}
+
+	existing := make(map[string]int)
+	for i, entry := range h.cfg.AmpCode.UpstreamAPIKeys {
+		existing[strings.TrimSpace(entry.UpstreamAPIKey)] = i
+	}
+
+	for _, newEntry := range body.Value {
+		upstreamKey := strings.TrimSpace(newEntry.UpstreamAPIKey)
+		if upstreamKey == "" {
+			continue
+		}
+		normalizedEntry := config.AmpUpstreamAPIKeyEntry{
+			UpstreamAPIKey: upstreamKey,
+			APIKeys:        normalizeAPIKeysList(newEntry.APIKeys),
+		}
+		if idx, ok := existing[upstreamKey]; ok {
+			h.cfg.AmpCode.UpstreamAPIKeys[idx] = normalizedEntry
+		} else {
+			h.cfg.AmpCode.UpstreamAPIKeys = append(h.cfg.AmpCode.UpstreamAPIKeys, normalizedEntry)
+			existing[upstreamKey] = len(h.cfg.AmpCode.UpstreamAPIKeys) - 1
+		}
+	}
+	h.persist(c)
+}
+
+// DeleteAmpUpstreamAPIKeys removes specified upstream API keys entries.
+// Body must be JSON: {"value": ["<upstream-api-key>", ...]}.
+// If "value" is an empty array, clears all entries.
+// If JSON is invalid or "value" is missing/null, returns 400 and does not persist any change.
+func (h *Handler) DeleteAmpUpstreamAPIKeys(c *gin.Context) {
+	var body struct {
+		Value []string `json:"value"`
+	}
+	if err := c.ShouldBindJSON(&body); err != nil {
+		c.JSON(400, gin.H{"error": "invalid body"})
+		return
+	}
+
+	if body.Value == nil {
+		c.JSON(400, gin.H{"error": "missing value"})
+		return
+	}
+
+	// Empty array means clear all
+	if len(body.Value) == 0 {
+		h.cfg.AmpCode.UpstreamAPIKeys = nil
+		h.persist(c)
+		return
+	}
+
+	toRemove := make(map[string]bool)
+	for _, key := range body.Value {
+		trimmed := strings.TrimSpace(key)
+		if trimmed == "" {
+			continue
+		}
+		toRemove[trimmed] = true
+	}
+	if len(toRemove) == 0 {
+		c.JSON(400, gin.H{"error": "empty value"})
+		return
+	}
+
+	newEntries := make([]config.AmpUpstreamAPIKeyEntry, 0, len(h.cfg.AmpCode.UpstreamAPIKeys))
+	for _, entry := range h.cfg.AmpCode.UpstreamAPIKeys {
+		if !toRemove[strings.TrimSpace(entry.UpstreamAPIKey)] {
+			newEntries = append(newEntries, entry)
+		}
+	}
+	h.cfg.AmpCode.UpstreamAPIKeys = newEntries
+	h.persist(c)
+}
+
+// normalizeAmpUpstreamAPIKeyEntries normalizes a list of upstream API key entries.
+func normalizeAmpUpstreamAPIKeyEntries(entries []config.AmpUpstreamAPIKeyEntry) []config.AmpUpstreamAPIKeyEntry {
+	if len(entries) == 0 {
+		return nil
+	}
+	out := make([]config.AmpUpstreamAPIKeyEntry, 0, len(entries))
+	for _, entry := range entries {
+		upstreamKey := strings.TrimSpace(entry.UpstreamAPIKey)
+		if upstreamKey == "" {
+			continue
+		}
+		apiKeys := normalizeAPIKeysList(entry.APIKeys)
+		out = append(out, config.AmpUpstreamAPIKeyEntry{
+			UpstreamAPIKey: upstreamKey,
+			APIKeys:        apiKeys,
+		})
+	}
+	if len(out) == 0 {
+		return nil
+	}
+	return out
+}
+
+// normalizeAPIKeysList trims and filters empty strings from a list of API keys.
+func normalizeAPIKeysList(keys []string) []string {
+	if len(keys) == 0 {
+		return nil
+	}
+	out := make([]string, 0, len(keys))
+	for _, k := range keys {
+		trimmed := strings.TrimSpace(k)
+		if trimmed != "" {
+			out = append(out, trimmed)
+		}
+	}
+	if len(out) == 0 {
+		return nil
+	}
+	return out
+}

internal/api/handlers/management/handler.go

@@ -59,6 +59,11 @@ func NewHandler(cfg *config.Config, configFilePath string, manager *coreauth.Man
 	}
 }
 
+// NewHandler creates a new management handler instance.
+func NewHandlerWithoutConfigFilePath(cfg *config.Config, manager *coreauth.Manager) *Handler {
+	return NewHandler(cfg, "", manager)
+}
+
 // SetConfig updates the in-memory config reference when the server hot-reloads.
 func (h *Handler) SetConfig(cfg *config.Config) { h.cfg = cfg }
 

internal/api/handlers/management/logs.go

@@ -209,6 +209,94 @@ func (h *Handler) GetRequestErrorLogs(c *gin.Context) {
 	c.JSON(http.StatusOK, gin.H{"files": files})
 }
 
+// GetRequestLogByID finds and downloads a request log file by its request ID.
+// The ID is matched against the suffix of log file names (format: *-{requestID}.log).
+func (h *Handler) GetRequestLogByID(c *gin.Context) {
+	if h == nil {
+		c.JSON(http.StatusInternalServerError, gin.H{"error": "handler unavailable"})
+		return
+	}
+	if h.cfg == nil {
+		c.JSON(http.StatusServiceUnavailable, gin.H{"error": "configuration unavailable"})
+		return
+	}
+
+	dir := h.logDirectory()
+	if strings.TrimSpace(dir) == "" {
+		c.JSON(http.StatusInternalServerError, gin.H{"error": "log directory not configured"})
+		return
+	}
+
+	requestID := strings.TrimSpace(c.Param("id"))
+	if requestID == "" {
+		requestID = strings.TrimSpace(c.Query("id"))
+	}
+	if requestID == "" {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "missing request ID"})
+		return
+	}
+	if strings.ContainsAny(requestID, "/\\") {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "invalid request ID"})
+		return
+	}
+
+	entries, err := os.ReadDir(dir)
+	if err != nil {
+		if os.IsNotExist(err) {
+			c.JSON(http.StatusNotFound, gin.H{"error": "log directory not found"})
+			return
+		}
+		c.JSON(http.StatusInternalServerError, gin.H{"error": fmt.Sprintf("failed to list log directory: %v", err)})
+		return
+	}
+
+	suffix := "-" + requestID + ".log"
+	var matchedFile string
+	for _, entry := range entries {
+		if entry.IsDir() {
+			continue
+		}
+		name := entry.Name()
+		if strings.HasSuffix(name, suffix) {
+			matchedFile = name
+			break
+		}
+	}
+
+	if matchedFile == "" {
+		c.JSON(http.StatusNotFound, gin.H{"error": "log file not found for the given request ID"})
+		return
+	}
+
+	dirAbs, errAbs := filepath.Abs(dir)
+	if errAbs != nil {
+		c.JSON(http.StatusInternalServerError, gin.H{"error": fmt.Sprintf("failed to resolve log directory: %v", errAbs)})
+		return
+	}
+	fullPath := filepath.Clean(filepath.Join(dirAbs, matchedFile))
+	prefix := dirAbs + string(os.PathSeparator)
+	if !strings.HasPrefix(fullPath, prefix) {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "invalid log file path"})
+		return
+	}
+
+	info, errStat := os.Stat(fullPath)
+	if errStat != nil {
+		if os.IsNotExist(errStat) {
+			c.JSON(http.StatusNotFound, gin.H{"error": "log file not found"})
+			return
+		}
+		c.JSON(http.StatusInternalServerError, gin.H{"error": fmt.Sprintf("failed to read log file: %v", errStat)})
+		return
+	}
+	if info.IsDir() {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "invalid log file"})
+		return
+	}
+
+	c.FileAttachment(fullPath, matchedFile)
+}
+
 // DownloadRequestErrorLog downloads a specific error request log file by name.
 func (h *Handler) DownloadRequestErrorLog(c *gin.Context) {
 	if h == nil {

internal/api/handlers/management/usage.go

@@ -1,12 +1,25 @@
 package management
 
 import (
+	"encoding/json"
 	"net/http"
+	"time"
 
 	"github.com/gin-gonic/gin"
 	"github.com/router-for-me/CLIProxyAPI/v6/internal/usage"
 )
 
+type usageExportPayload struct {
+	Version    int                      `json:"version"`
+	ExportedAt time.Time                `json:"exported_at"`
+	Usage      usage.StatisticsSnapshot `json:"usage"`
+}
+
+type usageImportPayload struct {
+	Version int                      `json:"version"`
+	Usage   usage.StatisticsSnapshot `json:"usage"`
+}
+
 // GetUsageStatistics returns the in-memory request statistics snapshot.
 func (h *Handler) GetUsageStatistics(c *gin.Context) {
 	var snapshot usage.StatisticsSnapshot
@@ -18,3 +31,49 @@ func (h *Handler) GetUsageStatistics(c *gin.Context) {
 		"failed_requests": snapshot.FailureCount,
 	})
 }
+
+// ExportUsageStatistics returns a complete usage snapshot for backup/migration.
+func (h *Handler) ExportUsageStatistics(c *gin.Context) {
+	var snapshot usage.StatisticsSnapshot
+	if h != nil && h.usageStats != nil {
+		snapshot = h.usageStats.Snapshot()
+	}
+	c.JSON(http.StatusOK, usageExportPayload{
+		Version:    1,
+		ExportedAt: time.Now().UTC(),
+		Usage:      snapshot,
+	})
+}
+
+// ImportUsageStatistics merges a previously exported usage snapshot into memory.
+func (h *Handler) ImportUsageStatistics(c *gin.Context) {
+	if h == nil || h.usageStats == nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "usage statistics unavailable"})
+		return
+	}
+
+	data, err := c.GetRawData()
+	if err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "failed to read request body"})
+		return
+	}
+
+	var payload usageImportPayload
+	if err := json.Unmarshal(data, &payload); err != nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "invalid json"})
+		return
+	}
+	if payload.Version != 0 && payload.Version != 1 {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "unsupported version"})
+		return
+	}
+
+	result := h.usageStats.MergeSnapshot(payload.Usage)
+	snapshot := h.usageStats.Snapshot()
+	c.JSON(http.StatusOK, gin.H{
+		"added":           result.Added,
+		"skipped":         result.Skipped,
+		"total_requests":  snapshot.TotalRequests,
+		"failed_requests": snapshot.FailureCount,
+	})
+}

internal/api/middleware/request_logging.go

@@ -98,10 +98,11 @@ func captureRequestInfo(c *gin.Context) (*RequestInfo, error) {
 	}
 
 	return &RequestInfo{
-		URL:     url,
-		Method:  method,
-		Headers: headers,
-		Body:    body,
+		URL:       url,
+		Method:    method,
+		Headers:   headers,
+		Body:      body,
+		RequestID: logging.GetGinRequestID(c),
 	}, nil
 }
 

internal/api/middleware/response_writer.go

@@ -15,10 +15,11 @@ import (
 
 // RequestInfo holds essential details of an incoming HTTP request for logging purposes.
 type RequestInfo struct {
-	URL     string              // URL is the request URL.
-	Method  string              // Method is the HTTP method (e.g., GET, POST).
-	Headers map[string][]string // Headers contains the request headers.
-	Body    []byte              // Body is the raw request body.
+	URL       string              // URL is the request URL.
+	Method    string              // Method is the HTTP method (e.g., GET, POST).
+	Headers   map[string][]string // Headers contains the request headers.
+	Body      []byte              // Body is the raw request body.
+	RequestID string              // RequestID is the unique identifier for the request.
 }
 
 // ResponseWriterWrapper wraps the standard gin.ResponseWriter to intercept and log response data.
@@ -149,6 +150,7 @@ func (w *ResponseWriterWrapper) WriteHeader(statusCode int) {
 			w.requestInfo.Method,
 			w.requestInfo.Headers,
 			w.requestInfo.Body,
+			w.requestInfo.RequestID,
 		)
 		if err == nil {
 			w.streamWriter = streamWriter
@@ -346,7 +348,7 @@ func (w *ResponseWriterWrapper) logRequest(statusCode int, headers map[string][]
 	}
 
 	if loggerWithOptions, ok := w.logger.(interface {
-		LogRequestWithOptions(string, string, map[string][]string, []byte, int, map[string][]string, []byte, []byte, []byte, []*interfaces.ErrorMessage, bool) error
+		LogRequestWithOptions(string, string, map[string][]string, []byte, int, map[string][]string, []byte, []byte, []byte, []*interfaces.ErrorMessage, bool, string) error
 	}); ok {
 		return loggerWithOptions.LogRequestWithOptions(
 			w.requestInfo.URL,
@@ -360,6 +362,7 @@ func (w *ResponseWriterWrapper) logRequest(statusCode int, headers map[string][]
 			apiResponseBody,
 			apiResponseErrors,
 			forceLog,
+			w.requestInfo.RequestID,
 		)
 	}
 
@@ -374,5 +377,6 @@ func (w *ResponseWriterWrapper) logRequest(statusCode int, headers map[string][]
 		apiRequestBody,
 		apiResponseBody,
 		apiResponseErrors,
+		w.requestInfo.RequestID,
 	)
 }

internal/api/modules/amp/amp.go

@@ -227,11 +227,20 @@ func (m *AmpModule) OnConfigUpdated(cfg *config.Config) error {
 			}
 		}
 
-		// Check API key change
+		// Check API key change (both default and per-client mappings)
 		apiKeyChanged := m.hasAPIKeyChanged(oldSettings, &newSettings)
-		if apiKeyChanged {
+		upstreamAPIKeysChanged := m.hasUpstreamAPIKeysChanged(oldSettings, &newSettings)
+		if apiKeyChanged || upstreamAPIKeysChanged {
 			if m.secretSource != nil {
-				if ms, ok := m.secretSource.(*MultiSourceSecret); ok {
+				if ms, ok := m.secretSource.(*MappedSecretSource); ok {
+					if apiKeyChanged {
+						ms.UpdateDefaultExplicitKey(newSettings.UpstreamAPIKey)
+						ms.InvalidateCache()
+					}
+					if upstreamAPIKeysChanged {
+						ms.UpdateMappings(newSettings.UpstreamAPIKeys)
+					}
+				} else if ms, ok := m.secretSource.(*MultiSourceSecret); ok {
 					ms.UpdateExplicitKey(newSettings.UpstreamAPIKey)
 					ms.InvalidateCache()
 				}
@@ -251,10 +260,22 @@ func (m *AmpModule) OnConfigUpdated(cfg *config.Config) error {
 
 func (m *AmpModule) enableUpstreamProxy(upstreamURL string, settings *config.AmpCode) error {
 	if m.secretSource == nil {
-		m.secretSource = NewMultiSourceSecret(settings.UpstreamAPIKey, 0 /* default 5min */)
+		// Create MultiSourceSecret as the default source, then wrap with MappedSecretSource
+		defaultSource := NewMultiSourceSecret(settings.UpstreamAPIKey, 0 /* default 5min */)
+		mappedSource := NewMappedSecretSource(defaultSource)
+		mappedSource.UpdateMappings(settings.UpstreamAPIKeys)
+		m.secretSource = mappedSource
+	} else if ms, ok := m.secretSource.(*MappedSecretSource); ok {
+		ms.UpdateDefaultExplicitKey(settings.UpstreamAPIKey)
+		ms.InvalidateCache()
+		ms.UpdateMappings(settings.UpstreamAPIKeys)
 	} else if ms, ok := m.secretSource.(*MultiSourceSecret); ok {
+		// Legacy path: wrap existing MultiSourceSecret with MappedSecretSource
 		ms.UpdateExplicitKey(settings.UpstreamAPIKey)
 		ms.InvalidateCache()
+		mappedSource := NewMappedSecretSource(ms)
+		mappedSource.UpdateMappings(settings.UpstreamAPIKeys)
+		m.secretSource = mappedSource
 	}
 
 	proxy, err := createReverseProxy(upstreamURL, m.secretSource)
@@ -279,16 +300,23 @@ func (m *AmpModule) hasModelMappingsChanged(old *config.AmpCode, new *config.Amp
 		return true
 	}
 
-	// Build map for efficient comparison
-	oldMap := make(map[string]string, len(old.ModelMappings))
+	// Build map for efficient and robust comparison
+	type mappingInfo struct {
+		to    string
+		regex bool
+	}
+	oldMap := make(map[string]mappingInfo, len(old.ModelMappings))
 	for _, mapping := range old.ModelMappings {
-		oldMap[strings.TrimSpace(mapping.From)] = strings.TrimSpace(mapping.To)
+		oldMap[strings.TrimSpace(mapping.From)] = mappingInfo{
+			to:    strings.TrimSpace(mapping.To),
+			regex: mapping.Regex,
+		}
 	}
 
 	for _, mapping := range new.ModelMappings {
 		from := strings.TrimSpace(mapping.From)
 		to := strings.TrimSpace(mapping.To)
-		if oldTo, exists := oldMap[from]; !exists || oldTo != to {
+		if oldVal, exists := oldMap[from]; !exists || oldVal.to != to || oldVal.regex != mapping.Regex {
 			return true
 		}
 	}
@@ -306,6 +334,66 @@ func (m *AmpModule) hasAPIKeyChanged(old *config.AmpCode, new *config.AmpCode) b
 	return oldKey != newKey
 }
 
+// hasUpstreamAPIKeysChanged compares old and new per-client upstream API key mappings.
+func (m *AmpModule) hasUpstreamAPIKeysChanged(old *config.AmpCode, new *config.AmpCode) bool {
+	if old == nil {
+		return len(new.UpstreamAPIKeys) > 0
+	}
+
+	if len(old.UpstreamAPIKeys) != len(new.UpstreamAPIKeys) {
+		return true
+	}
+
+	// Build map for comparison: upstreamKey -> set of clientKeys
+	type entryInfo struct {
+		upstreamKey string
+		clientKeys  map[string]struct{}
+	}
+	oldEntries := make([]entryInfo, len(old.UpstreamAPIKeys))
+	for i, entry := range old.UpstreamAPIKeys {
+		clientKeys := make(map[string]struct{}, len(entry.APIKeys))
+		for _, k := range entry.APIKeys {
+			trimmed := strings.TrimSpace(k)
+			if trimmed == "" {
+				continue
+			}
+			clientKeys[trimmed] = struct{}{}
+		}
+		oldEntries[i] = entryInfo{
+			upstreamKey: strings.TrimSpace(entry.UpstreamAPIKey),
+			clientKeys:  clientKeys,
+		}
+	}
+
+	for i, newEntry := range new.UpstreamAPIKeys {
+		if i >= len(oldEntries) {
+			return true
+		}
+		oldE := oldEntries[i]
+		if strings.TrimSpace(newEntry.UpstreamAPIKey) != oldE.upstreamKey {
+			return true
+		}
+		newKeys := make(map[string]struct{}, len(newEntry.APIKeys))
+		for _, k := range newEntry.APIKeys {
+			trimmed := strings.TrimSpace(k)
+			if trimmed == "" {
+				continue
+			}
+			newKeys[trimmed] = struct{}{}
+		}
+		if len(newKeys) != len(oldE.clientKeys) {
+			return true
+		}
+		for k := range newKeys {
+			if _, ok := oldE.clientKeys[k]; !ok {
+				return true
+			}
+		}
+	}
+
+	return false
+}
+
 // GetModelMapper returns the model mapper instance (for testing/debugging).
 func (m *AmpModule) GetModelMapper() *DefaultModelMapper {
 	return m.modelMapper

internal/api/modules/amp/amp_test.go

@@ -312,3 +312,41 @@ func TestAmpModule_ProviderAliasesAlwaysRegistered(t *testing.T) {
 		})
 	}
 }
+
+func TestAmpModule_hasUpstreamAPIKeysChanged_DetectsRemovedKeyWithDuplicateInput(t *testing.T) {
+	m := &AmpModule{}
+
+	oldCfg := &config.AmpCode{
+		UpstreamAPIKeys: []config.AmpUpstreamAPIKeyEntry{
+			{UpstreamAPIKey: "u1", APIKeys: []string{"k1", "k2"}},
+		},
+	}
+	newCfg := &config.AmpCode{
+		UpstreamAPIKeys: []config.AmpUpstreamAPIKeyEntry{
+			{UpstreamAPIKey: "u1", APIKeys: []string{"k1", "k1"}},
+		},
+	}
+
+	if !m.hasUpstreamAPIKeysChanged(oldCfg, newCfg) {
+		t.Fatal("expected change to be detected when k2 is removed but new list contains duplicates")
+	}
+}
+
+func TestAmpModule_hasUpstreamAPIKeysChanged_IgnoresEmptyAndWhitespaceKeys(t *testing.T) {
+	m := &AmpModule{}
+
+	oldCfg := &config.AmpCode{
+		UpstreamAPIKeys: []config.AmpUpstreamAPIKeyEntry{
+			{UpstreamAPIKey: "u1", APIKeys: []string{"k1", "k2"}},
+		},
+	}
+	newCfg := &config.AmpCode{
+		UpstreamAPIKeys: []config.AmpUpstreamAPIKeyEntry{
+			{UpstreamAPIKey: "u1", APIKeys: []string{"  k1  ", "", "k2", "   "}},
+		},
+	}
+
+	if m.hasUpstreamAPIKeysChanged(oldCfg, newCfg) {
+		t.Fatal("expected no change when only whitespace/empty entries differ")
+	}
+}

internal/api/modules/amp/model_mapping.go

@@ -3,6 +3,7 @@
 package amp
 
 import (
+	"regexp"
 	"strings"
 	"sync"
 
@@ -26,13 +27,15 @@ type ModelMapper interface {
 // DefaultModelMapper implements ModelMapper with thread-safe mapping storage.
 type DefaultModelMapper struct {
 	mu       sync.RWMutex
-	mappings map[string]string // from -> to (normalized lowercase keys)
+	mappings map[string]string // exact: from -> to (normalized lowercase keys)
+	regexps  []regexMapping    // regex rules evaluated in order
 }
 
 // NewModelMapper creates a new model mapper with the given initial mappings.
 func NewModelMapper(mappings []config.AmpModelMapping) *DefaultModelMapper {
 	m := &DefaultModelMapper{
 		mappings: make(map[string]string),
+		regexps:  nil,
 	}
 	m.UpdateMappings(mappings)
 	return m
@@ -55,7 +58,18 @@ func (m *DefaultModelMapper) MapModel(requestedModel string) string {
 	// Check for direct mapping
 	targetModel, exists := m.mappings[normalizedRequest]
 	if !exists {
-		return ""
+		// Try regex mappings in order
+		base, _ := util.NormalizeThinkingModel(requestedModel)
+		for _, rm := range m.regexps {
+			if rm.re.MatchString(requestedModel) || (base != "" && rm.re.MatchString(base)) {
+				targetModel = rm.to
+				exists = true
+				break
+			}
+		}
+		if !exists {
+			return ""
+		}
 	}
 
 	// Verify target model has available providers
@@ -78,6 +92,7 @@ func (m *DefaultModelMapper) UpdateMappings(mappings []config.AmpModelMapping) {
 
 	// Clear and rebuild mappings
 	m.mappings = make(map[string]string, len(mappings))
+	m.regexps = make([]regexMapping, 0, len(mappings))
 
 	for _, mapping := range mappings {
 		from := strings.TrimSpace(mapping.From)
@@ -88,16 +103,30 @@ func (m *DefaultModelMapper) UpdateMappings(mappings []config.AmpModelMapping) {
 			continue
 		}
 
-		// Store with normalized lowercase key for case-insensitive lookup
-		normalizedFrom := strings.ToLower(from)
-		m.mappings[normalizedFrom] = to
-
-		log.Debugf("amp model mapping registered: %s -> %s", from, to)
+		if mapping.Regex {
+			// Compile case-insensitive regex; wrap with (?i) to match behavior of exact lookups
+			pattern := "(?i)" + from
+			re, err := regexp.Compile(pattern)
+			if err != nil {
+				log.Warnf("amp model mapping: invalid regex %q: %v", from, err)
+				continue
+			}
+			m.regexps = append(m.regexps, regexMapping{re: re, to: to})
+			log.Debugf("amp model regex mapping registered: /%s/ -> %s", from, to)
+		} else {
+			// Store with normalized lowercase key for case-insensitive lookup
+			normalizedFrom := strings.ToLower(from)
+			m.mappings[normalizedFrom] = to
+			log.Debugf("amp model mapping registered: %s -> %s", from, to)
+		}
 	}
 
 	if len(m.mappings) > 0 {
 		log.Infof("amp model mapping: loaded %d mapping(s)", len(m.mappings))
 	}
+	if n := len(m.regexps); n > 0 {
+		log.Infof("amp model mapping: loaded %d regex mapping(s)", n)
+	}
 }
 
 // GetMappings returns a copy of current mappings (for debugging/status).
@@ -111,3 +140,8 @@ func (m *DefaultModelMapper) GetMappings() map[string]string {
 	}
 	return result
 }
+
+type regexMapping struct {
+	re *regexp.Regexp
+	to string
+}

internal/api/modules/amp/model_mapping_test.go

@@ -203,3 +203,81 @@ func TestModelMapper_GetMappings_ReturnsCopy(t *testing.T) {
 		t.Error("Original map was modified")
 	}
 }
+
+func TestModelMapper_Regex_MatchBaseWithoutParens(t *testing.T) {
+	reg := registry.GetGlobalRegistry()
+	reg.RegisterClient("test-client-regex-1", "gemini", []*registry.ModelInfo{
+		{ID: "gemini-2.5-pro", OwnedBy: "google", Type: "gemini"},
+	})
+	defer reg.UnregisterClient("test-client-regex-1")
+
+	mappings := []config.AmpModelMapping{
+		{From: "^gpt-5$", To: "gemini-2.5-pro", Regex: true},
+	}
+
+	mapper := NewModelMapper(mappings)
+
+	// Incoming model has reasoning suffix but should match base via regex
+	result := mapper.MapModel("gpt-5(high)")
+	if result != "gemini-2.5-pro" {
+		t.Errorf("Expected gemini-2.5-pro, got %s", result)
+	}
+}
+
+func TestModelMapper_Regex_ExactPrecedence(t *testing.T) {
+	reg := registry.GetGlobalRegistry()
+	reg.RegisterClient("test-client-regex-2", "claude", []*registry.ModelInfo{
+		{ID: "claude-sonnet-4", OwnedBy: "anthropic", Type: "claude"},
+	})
+	reg.RegisterClient("test-client-regex-3", "gemini", []*registry.ModelInfo{
+		{ID: "gemini-2.5-pro", OwnedBy: "google", Type: "gemini"},
+	})
+	defer reg.UnregisterClient("test-client-regex-2")
+	defer reg.UnregisterClient("test-client-regex-3")
+
+	mappings := []config.AmpModelMapping{
+		{From: "gpt-5", To: "claude-sonnet-4"},                 // exact
+		{From: "^gpt-5.*$", To: "gemini-2.5-pro", Regex: true}, // regex
+	}
+
+	mapper := NewModelMapper(mappings)
+
+	// Exact match should win over regex
+	result := mapper.MapModel("gpt-5")
+	if result != "claude-sonnet-4" {
+		t.Errorf("Expected claude-sonnet-4, got %s", result)
+	}
+}
+
+func TestModelMapper_Regex_InvalidPattern_Skipped(t *testing.T) {
+	// Invalid regex should be skipped and not cause panic
+	mappings := []config.AmpModelMapping{
+		{From: "(", To: "target", Regex: true},
+	}
+
+	mapper := NewModelMapper(mappings)
+
+	result := mapper.MapModel("anything")
+	if result != "" {
+		t.Errorf("Expected empty result due to invalid regex, got %s", result)
+	}
+}
+
+func TestModelMapper_Regex_CaseInsensitive(t *testing.T) {
+	reg := registry.GetGlobalRegistry()
+	reg.RegisterClient("test-client-regex-4", "claude", []*registry.ModelInfo{
+		{ID: "claude-sonnet-4", OwnedBy: "anthropic", Type: "claude"},
+	})
+	defer reg.UnregisterClient("test-client-regex-4")
+
+	mappings := []config.AmpModelMapping{
+		{From: "^CLAUDE-OPUS-.*$", To: "claude-sonnet-4", Regex: true},
+	}
+
+	mapper := NewModelMapper(mappings)
+
+	result := mapper.MapModel("claude-opus-4.5")
+	if result != "claude-sonnet-4" {
+		t.Errorf("Expected claude-sonnet-4, got %s", result)
+	}
+}

internal/api/modules/amp/proxy.go

@@ -18,6 +18,33 @@ import (
 	log "github.com/sirupsen/logrus"
 )
 
+func removeQueryValuesMatching(req *http.Request, key string, match string) {
+	if req == nil || req.URL == nil || match == "" {
+		return
+	}
+
+	q := req.URL.Query()
+	values, ok := q[key]
+	if !ok || len(values) == 0 {
+		return
+	}
+
+	kept := make([]string, 0, len(values))
+	for _, v := range values {
+		if v == match {
+			continue
+		}
+		kept = append(kept, v)
+	}
+
+	if len(kept) == 0 {
+		q.Del(key)
+	} else {
+		q[key] = kept
+	}
+	req.URL.RawQuery = q.Encode()
+}
+
 // readCloser wraps a reader and forwards Close to a separate closer.
 // Used to restore peeked bytes while preserving upstream body Close behavior.
 type readCloser struct {
@@ -48,6 +75,14 @@ func createReverseProxy(upstreamURL string, secretSource SecretSource) (*httputi
 		// We will set our own Authorization using the configured upstream-api-key
 		req.Header.Del("Authorization")
 		req.Header.Del("X-Api-Key")
+		req.Header.Del("X-Goog-Api-Key")
+
+		// Remove query-based credentials if they match the authenticated client API key.
+		// This prevents leaking client auth material to the Amp upstream while avoiding
+		// breaking unrelated upstream query parameters.
+		clientKey := getClientAPIKeyFromContext(req.Context())
+		removeQueryValuesMatching(req, "key", clientKey)
+		removeQueryValuesMatching(req, "auth_token", clientKey)
 
 		// Preserve correlation headers for debugging
 		if req.Header.Get("X-Request-ID") == "" {

internal/api/modules/amp/proxy_test.go

@@ -3,11 +3,15 @@ package amp
 import (
 	"bytes"
 	"compress/gzip"
+	"context"
 	"fmt"
 	"io"
 	"net/http"
 	"net/http/httptest"
+	"strings"
 	"testing"
+
+	"github.com/router-for-me/CLIProxyAPI/v6/internal/config"
 )
 
 // Helper: compress data with gzip
@@ -306,6 +310,159 @@ func TestReverseProxy_EmptySecret(t *testing.T) {
 	}
 }
 
+func TestReverseProxy_StripsClientCredentialsFromHeadersAndQuery(t *testing.T) {
+	type captured struct {
+		headers http.Header
+		query   string
+	}
+	got := make(chan captured, 1)
+	upstream := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		got <- captured{headers: r.Header.Clone(), query: r.URL.RawQuery}
+		w.WriteHeader(200)
+		w.Write([]byte(`ok`))
+	}))
+	defer upstream.Close()
+
+	proxy, err := createReverseProxy(upstream.URL, NewStaticSecretSource("upstream"))
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		// Simulate clientAPIKeyMiddleware injection (per-request)
+		ctx := context.WithValue(r.Context(), clientAPIKeyContextKey{}, "client-key")
+		proxy.ServeHTTP(w, r.WithContext(ctx))
+	}))
+	defer srv.Close()
+
+	req, err := http.NewRequest(http.MethodGet, srv.URL+"/test?key=client-key&key=keep&auth_token=client-key&foo=bar", nil)
+	if err != nil {
+		t.Fatal(err)
+	}
+	req.Header.Set("Authorization", "Bearer client-key")
+	req.Header.Set("X-Api-Key", "client-key")
+	req.Header.Set("X-Goog-Api-Key", "client-key")
+
+	res, err := http.DefaultClient.Do(req)
+	if err != nil {
+		t.Fatal(err)
+	}
+	res.Body.Close()
+
+	c := <-got
+
+	// These are client-provided credentials and must not reach the upstream.
+	if v := c.headers.Get("X-Goog-Api-Key"); v != "" {
+		t.Fatalf("X-Goog-Api-Key should be stripped, got: %q", v)
+	}
+
+	// We inject upstream Authorization/X-Api-Key, so the client auth must not survive.
+	if v := c.headers.Get("Authorization"); v != "Bearer upstream" {
+		t.Fatalf("Authorization should be upstream-injected, got: %q", v)
+	}
+	if v := c.headers.Get("X-Api-Key"); v != "upstream" {
+		t.Fatalf("X-Api-Key should be upstream-injected, got: %q", v)
+	}
+
+	// Query-based credentials should be stripped only when they match the authenticated client key.
+	// Should keep unrelated values and parameters.
+	if strings.Contains(c.query, "auth_token=client-key") || strings.Contains(c.query, "key=client-key") {
+		t.Fatalf("query credentials should be stripped, got raw query: %q", c.query)
+	}
+	if !strings.Contains(c.query, "key=keep") || !strings.Contains(c.query, "foo=bar") {
+		t.Fatalf("expected query to keep non-credential params, got raw query: %q", c.query)
+	}
+}
+
+func TestReverseProxy_InjectsMappedSecret_FromRequestContext(t *testing.T) {
+	gotHeaders := make(chan http.Header, 1)
+	upstream := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		gotHeaders <- r.Header.Clone()
+		w.WriteHeader(200)
+		w.Write([]byte(`ok`))
+	}))
+	defer upstream.Close()
+
+	defaultSource := NewStaticSecretSource("default")
+	mapped := NewMappedSecretSource(defaultSource)
+	mapped.UpdateMappings([]config.AmpUpstreamAPIKeyEntry{
+		{
+			UpstreamAPIKey: "u1",
+			APIKeys:        []string{"k1"},
+		},
+	})
+
+	proxy, err := createReverseProxy(upstream.URL, mapped)
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		// Simulate clientAPIKeyMiddleware injection (per-request)
+		ctx := context.WithValue(r.Context(), clientAPIKeyContextKey{}, "k1")
+		proxy.ServeHTTP(w, r.WithContext(ctx))
+	}))
+	defer srv.Close()
+
+	res, err := http.Get(srv.URL + "/test")
+	if err != nil {
+		t.Fatal(err)
+	}
+	res.Body.Close()
+
+	hdr := <-gotHeaders
+	if hdr.Get("X-Api-Key") != "u1" {
+		t.Fatalf("X-Api-Key missing or wrong, got: %q", hdr.Get("X-Api-Key"))
+	}
+	if hdr.Get("Authorization") != "Bearer u1" {
+		t.Fatalf("Authorization missing or wrong, got: %q", hdr.Get("Authorization"))
+	}
+}
+
+func TestReverseProxy_MappedSecret_FallsBackToDefault(t *testing.T) {
+	gotHeaders := make(chan http.Header, 1)
+	upstream := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		gotHeaders <- r.Header.Clone()
+		w.WriteHeader(200)
+		w.Write([]byte(`ok`))
+	}))
+	defer upstream.Close()
+
+	defaultSource := NewStaticSecretSource("default")
+	mapped := NewMappedSecretSource(defaultSource)
+	mapped.UpdateMappings([]config.AmpUpstreamAPIKeyEntry{
+		{
+			UpstreamAPIKey: "u1",
+			APIKeys:        []string{"k1"},
+		},
+	})
+
+	proxy, err := createReverseProxy(upstream.URL, mapped)
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	srv := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		ctx := context.WithValue(r.Context(), clientAPIKeyContextKey{}, "k2")
+		proxy.ServeHTTP(w, r.WithContext(ctx))
+	}))
+	defer srv.Close()
+
+	res, err := http.Get(srv.URL + "/test")
+	if err != nil {
+		t.Fatal(err)
+	}
+	res.Body.Close()
+
+	hdr := <-gotHeaders
+	if hdr.Get("X-Api-Key") != "default" {
+		t.Fatalf("X-Api-Key fallback missing or wrong, got: %q", hdr.Get("X-Api-Key"))
+	}
+	if hdr.Get("Authorization") != "Bearer default" {
+		t.Fatalf("Authorization fallback missing or wrong, got: %q", hdr.Get("Authorization"))
+	}
+}
+
 func TestReverseProxy_ErrorHandler(t *testing.T) {
 	// Point proxy to a non-routable address to trigger error
 	proxy, err := createReverseProxy("http://127.0.0.1:1", NewStaticSecretSource(""))

internal/api/modules/amp/routes.go

@@ -1,6 +1,7 @@
 package amp
 
 import (
+	"context"
 	"errors"
 	"net"
 	"net/http"
@@ -16,6 +17,37 @@ import (
 	log "github.com/sirupsen/logrus"
 )
 
+// clientAPIKeyContextKey is the context key used to pass the client API key
+// from gin.Context to the request context for SecretSource lookup.
+type clientAPIKeyContextKey struct{}
+
+// clientAPIKeyMiddleware injects the authenticated client API key from gin.Context["apiKey"]
+// into the request context so that SecretSource can look it up for per-client upstream routing.
+func clientAPIKeyMiddleware() gin.HandlerFunc {
+	return func(c *gin.Context) {
+		// Extract the client API key from gin context (set by AuthMiddleware)
+		if apiKey, exists := c.Get("apiKey"); exists {
+			if keyStr, ok := apiKey.(string); ok && keyStr != "" {
+				// Inject into request context for SecretSource.Get(ctx) to read
+				ctx := context.WithValue(c.Request.Context(), clientAPIKeyContextKey{}, keyStr)
+				c.Request = c.Request.WithContext(ctx)
+			}
+		}
+		c.Next()
+	}
+}
+
+// getClientAPIKeyFromContext retrieves the client API key from request context.
+// Returns empty string if not present.
+func getClientAPIKeyFromContext(ctx context.Context) string {
+	if val := ctx.Value(clientAPIKeyContextKey{}); val != nil {
+		if keyStr, ok := val.(string); ok {
+			return keyStr
+		}
+	}
+	return ""
+}
+
 // localhostOnlyMiddleware returns a middleware that dynamically checks the module's
 // localhost restriction setting. This allows hot-reload of the restriction without restarting.
 func (m *AmpModule) localhostOnlyMiddleware() gin.HandlerFunc {
@@ -129,6 +161,9 @@ func (m *AmpModule) registerManagementRoutes(engine *gin.Engine, baseHandler *ha
 		authWithBypass = wrapManagementAuth(auth, "/threads", "/auth", "/docs", "/settings")
 	}
 
+	// Inject client API key into request context for per-client upstream routing
+	ampAPI.Use(clientAPIKeyMiddleware())
+
 	// Dynamic proxy handler that uses m.getProxy() for hot-reload support
 	proxyHandler := func(c *gin.Context) {
 		// Swallow ErrAbortHandler panics from ReverseProxy copyResponse to avoid noisy stack traces
@@ -175,6 +210,8 @@ func (m *AmpModule) registerManagementRoutes(engine *gin.Engine, baseHandler *ha
 	if authWithBypass != nil {
 		rootMiddleware = append(rootMiddleware, authWithBypass)
 	}
+	// Add clientAPIKeyMiddleware after auth for per-client upstream routing
+	rootMiddleware = append(rootMiddleware, clientAPIKeyMiddleware())
 	engine.GET("/threads", append(rootMiddleware, proxyHandler)...)
 	engine.GET("/threads/*path", append(rootMiddleware, proxyHandler)...)
 	engine.GET("/docs", append(rootMiddleware, proxyHandler)...)
@@ -244,6 +281,8 @@ func (m *AmpModule) registerProviderAliases(engine *gin.Engine, baseHandler *han
 	if auth != nil {
 		ampProviders.Use(auth)
 	}
+	// Inject client API key into request context for per-client upstream routing
+	ampProviders.Use(clientAPIKeyMiddleware())
 
 	provider := ampProviders.Group("/:provider")
 

internal/api/modules/amp/secret.go

@@ -9,6 +9,9 @@ import (
 	"strings"
 	"sync"
 	"time"
+
+	"github.com/router-for-me/CLIProxyAPI/v6/internal/config"
+	log "github.com/sirupsen/logrus"
 )
 
 // SecretSource provides Amp API keys with configurable precedence and caching
@@ -164,3 +167,82 @@ func NewStaticSecretSource(key string) *StaticSecretSource {
 func (s *StaticSecretSource) Get(ctx context.Context) (string, error) {
 	return s.key, nil
 }
+
+// MappedSecretSource wraps a default SecretSource and adds per-client API key mapping.
+// When a request context contains a client API key that matches a configured mapping,
+// the corresponding upstream key is returned. Otherwise, falls back to the default source.
+type MappedSecretSource struct {
+	defaultSource SecretSource
+	mu            sync.RWMutex
+	lookup        map[string]string // clientKey -> upstreamKey
+}
+
+// NewMappedSecretSource creates a MappedSecretSource wrapping the given default source.
+func NewMappedSecretSource(defaultSource SecretSource) *MappedSecretSource {
+	return &MappedSecretSource{
+		defaultSource: defaultSource,
+		lookup:        make(map[string]string),
+	}
+}
+
+// Get retrieves the Amp API key, checking per-client mappings first.
+// If the request context contains a client API key that matches a configured mapping,
+// returns the corresponding upstream key. Otherwise, falls back to the default source.
+func (s *MappedSecretSource) Get(ctx context.Context) (string, error) {
+	// Try to get client API key from request context
+	clientKey := getClientAPIKeyFromContext(ctx)
+	if clientKey != "" {
+		s.mu.RLock()
+		if upstreamKey, ok := s.lookup[clientKey]; ok && upstreamKey != "" {
+			s.mu.RUnlock()
+			return upstreamKey, nil
+		}
+		s.mu.RUnlock()
+	}
+
+	// Fall back to default source
+	return s.defaultSource.Get(ctx)
+}
+
+// UpdateMappings rebuilds the client-to-upstream key mapping from configuration entries.
+// If the same client key appears in multiple entries, logs a warning and uses the first one.
+func (s *MappedSecretSource) UpdateMappings(entries []config.AmpUpstreamAPIKeyEntry) {
+	newLookup := make(map[string]string)
+
+	for _, entry := range entries {
+		upstreamKey := strings.TrimSpace(entry.UpstreamAPIKey)
+		if upstreamKey == "" {
+			continue
+		}
+		for _, clientKey := range entry.APIKeys {
+			trimmedKey := strings.TrimSpace(clientKey)
+			if trimmedKey == "" {
+				continue
+			}
+			if _, exists := newLookup[trimmedKey]; exists {
+				// Log warning for duplicate client key, first one wins
+				log.Warnf("amp upstream-api-keys: client API key appears in multiple entries; using first mapping.")
+				continue
+			}
+			newLookup[trimmedKey] = upstreamKey
+		}
+	}
+
+	s.mu.Lock()
+	s.lookup = newLookup
+	s.mu.Unlock()
+}
+
+// UpdateDefaultExplicitKey updates the explicit key on the underlying MultiSourceSecret (if applicable).
+func (s *MappedSecretSource) UpdateDefaultExplicitKey(key string) {
+	if ms, ok := s.defaultSource.(*MultiSourceSecret); ok {
+		ms.UpdateExplicitKey(key)
+	}
+}
+
+// InvalidateCache invalidates cache on the underlying MultiSourceSecret (if applicable).
+func (s *MappedSecretSource) InvalidateCache() {
+	if ms, ok := s.defaultSource.(*MultiSourceSecret); ok {
+		ms.InvalidateCache()
+	}
+}

internal/api/modules/amp/secret_test.go

@@ -8,6 +8,10 @@ import (
 	"sync"
 	"testing"
 	"time"
+
+	"github.com/router-for-me/CLIProxyAPI/v6/internal/config"
+	log "github.com/sirupsen/logrus"
+	"github.com/sirupsen/logrus/hooks/test"
 )
 
 func TestMultiSourceSecret_PrecedenceOrder(t *testing.T) {
@@ -278,3 +282,85 @@ func TestMultiSourceSecret_CacheEmptyResult(t *testing.T) {
 		t.Fatalf("after cache expiry, expected new-value, got %q", got3)
 	}
 }
+
+func TestMappedSecretSource_UsesMappingFromContext(t *testing.T) {
+	defaultSource := NewStaticSecretSource("default")
+	s := NewMappedSecretSource(defaultSource)
+	s.UpdateMappings([]config.AmpUpstreamAPIKeyEntry{
+		{
+			UpstreamAPIKey: "u1",
+			APIKeys:        []string{"k1"},
+		},
+	})
+
+	ctx := context.WithValue(context.Background(), clientAPIKeyContextKey{}, "k1")
+	got, err := s.Get(ctx)
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+	if got != "u1" {
+		t.Fatalf("want u1, got %q", got)
+	}
+
+	ctx = context.WithValue(context.Background(), clientAPIKeyContextKey{}, "k2")
+	got, err = s.Get(ctx)
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+	if got != "default" {
+		t.Fatalf("want default fallback, got %q", got)
+	}
+}
+
+func TestMappedSecretSource_DuplicateClientKey_FirstWins(t *testing.T) {
+	defaultSource := NewStaticSecretSource("default")
+	s := NewMappedSecretSource(defaultSource)
+	s.UpdateMappings([]config.AmpUpstreamAPIKeyEntry{
+		{
+			UpstreamAPIKey: "u1",
+			APIKeys:        []string{"k1"},
+		},
+		{
+			UpstreamAPIKey: "u2",
+			APIKeys:        []string{"k1"},
+		},
+	})
+
+	ctx := context.WithValue(context.Background(), clientAPIKeyContextKey{}, "k1")
+	got, err := s.Get(ctx)
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+	if got != "u1" {
+		t.Fatalf("want u1 (first wins), got %q", got)
+	}
+}
+
+func TestMappedSecretSource_DuplicateClientKey_LogsWarning(t *testing.T) {
+	hook := test.NewLocal(log.StandardLogger())
+	defer hook.Reset()
+
+	defaultSource := NewStaticSecretSource("default")
+	s := NewMappedSecretSource(defaultSource)
+	s.UpdateMappings([]config.AmpUpstreamAPIKeyEntry{
+		{
+			UpstreamAPIKey: "u1",
+			APIKeys:        []string{"k1"},
+		},
+		{
+			UpstreamAPIKey: "u2",
+			APIKeys:        []string{"k1"},
+		},
+	})
+
+	foundWarning := false
+	for _, entry := range hook.AllEntries() {
+		if entry.Level == log.WarnLevel && entry.Message == "amp upstream-api-keys: client API key appears in multiple entries; using first mapping." {
+			foundWarning = true
+			break
+		}
+	}
+	if !foundWarning {
+		t.Fatal("expected warning log for duplicate client key, but none was found")
+	}
+}

internal/api/server.go

@@ -209,13 +209,15 @@ func NewServer(cfg *config.Config, authManager *auth.Manager, accessManager *sdk
 	// Resolve logs directory relative to the configuration file directory.
 	var requestLogger logging.RequestLogger
 	var toggle func(bool)
-	if optionState.requestLoggerFactory != nil {
-		requestLogger = optionState.requestLoggerFactory(cfg, configFilePath)
-	}
-	if requestLogger != nil {
-		engine.Use(middleware.RequestLoggingMiddleware(requestLogger))
-		if setter, ok := requestLogger.(interface{ SetEnabled(bool) }); ok {
-			toggle = setter.SetEnabled
+	if !cfg.CommercialMode {
+		if optionState.requestLoggerFactory != nil {
+			requestLogger = optionState.requestLoggerFactory(cfg, configFilePath)
+		}
+		if requestLogger != nil {
+			engine.Use(middleware.RequestLoggingMiddleware(requestLogger))
+			if setter, ok := requestLogger.(interface{ SetEnabled(bool) }); ok {
+				toggle = setter.SetEnabled
+			}
 		}
 	}
 
@@ -494,6 +496,8 @@ func (s *Server) registerManagementRoutes() {
 	mgmt.Use(s.managementAvailabilityMiddleware(), s.mgmt.Middleware())
 	{
 		mgmt.GET("/usage", s.mgmt.GetUsageStatistics)
+		mgmt.GET("/usage/export", s.mgmt.ExportUsageStatistics)
+		mgmt.POST("/usage/import", s.mgmt.ImportUsageStatistics)
 		mgmt.GET("/config", s.mgmt.GetConfig)
 		mgmt.GET("/config.yaml", s.mgmt.GetConfigYAML)
 		mgmt.PUT("/config.yaml", s.mgmt.PutConfigYAML)
@@ -516,6 +520,8 @@ func (s *Server) registerManagementRoutes() {
 		mgmt.PATCH("/proxy-url", s.mgmt.PutProxyURL)
 		mgmt.DELETE("/proxy-url", s.mgmt.DeleteProxyURL)
 
+		mgmt.POST("/api-call", s.mgmt.APICall)
+
 		mgmt.GET("/quota-exceeded/switch-project", s.mgmt.GetSwitchProject)
 		mgmt.PUT("/quota-exceeded/switch-project", s.mgmt.PutSwitchProject)
 		mgmt.PATCH("/quota-exceeded/switch-project", s.mgmt.PutSwitchProject)
@@ -538,6 +544,7 @@ func (s *Server) registerManagementRoutes() {
 		mgmt.DELETE("/logs", s.mgmt.DeleteLogs)
 		mgmt.GET("/request-error-logs", s.mgmt.GetRequestErrorLogs)
 		mgmt.GET("/request-error-logs/:name", s.mgmt.DownloadRequestErrorLog)
+		mgmt.GET("/request-log-by-id/:id", s.mgmt.GetRequestLogByID)
 		mgmt.GET("/request-log", s.mgmt.GetRequestLog)
 		mgmt.PUT("/request-log", s.mgmt.PutRequestLog)
 		mgmt.PATCH("/request-log", s.mgmt.PutRequestLog)
@@ -564,6 +571,10 @@ func (s *Server) registerManagementRoutes() {
 		mgmt.GET("/ampcode/force-model-mappings", s.mgmt.GetAmpForceModelMappings)
 		mgmt.PUT("/ampcode/force-model-mappings", s.mgmt.PutAmpForceModelMappings)
 		mgmt.PATCH("/ampcode/force-model-mappings", s.mgmt.PutAmpForceModelMappings)
+		mgmt.GET("/ampcode/upstream-api-keys", s.mgmt.GetAmpUpstreamAPIKeys)
+		mgmt.PUT("/ampcode/upstream-api-keys", s.mgmt.PutAmpUpstreamAPIKeys)
+		mgmt.PATCH("/ampcode/upstream-api-keys", s.mgmt.PatchAmpUpstreamAPIKeys)
+		mgmt.DELETE("/ampcode/upstream-api-keys", s.mgmt.DeleteAmpUpstreamAPIKeys)
 
 		mgmt.GET("/request-retry", s.mgmt.GetRequestRetry)
 		mgmt.PUT("/request-retry", s.mgmt.PutRequestRetry)
@@ -866,7 +877,7 @@ func (s *Server) UpdateClients(cfg *config.Config) {
 	}
 
 	if oldCfg == nil || oldCfg.LoggingToFile != cfg.LoggingToFile || oldCfg.LogsMaxTotalSizeMB != cfg.LogsMaxTotalSizeMB {
-		if err := logging.ConfigureLogOutput(cfg.LoggingToFile, cfg.LogsMaxTotalSizeMB); err != nil {
+		if err := logging.ConfigureLogOutput(cfg); err != nil {
 			log.Errorf("failed to reconfigure log output: %v", err)
 		} else {
 			if oldCfg == nil {

internal/auth/kiro/aws.go

@@ -40,6 +40,10 @@ type KiroTokenData struct {
 	ClientSecret string `json:"clientSecret,omitempty"`
 	// Email is the user's email address (used for file naming)
 	Email string `json:"email,omitempty"`
+	// StartURL is the IDC/Identity Center start URL (only for IDC auth method)
+	StartURL string `json:"startUrl,omitempty"`
+	// Region is the AWS region for IDC authentication (only for IDC auth method)
+	Region string `json:"region,omitempty"`
 }
 
 // KiroAuthBundle aggregates authentication data after OAuth flow completion

internal/auth/kiro/sso_oidc.go

@@ -2,16 +2,19 @@
 package kiro
 
 import (
+	"bufio"
 	"context"
 	"crypto/rand"
 	"crypto/sha256"
 	"encoding/base64"
 	"encoding/json"
+	"errors"
 	"fmt"
 	"html"
 	"io"
 	"net"
 	"net/http"
+	"os"
 	"strings"
 	"time"
 
@@ -24,19 +27,31 @@ import (
 const (
 	// AWS SSO OIDC endpoints
 	ssoOIDCEndpoint = "https://oidc.us-east-1.amazonaws.com"
-	
+
 	// Kiro's start URL for Builder ID
 	builderIDStartURL = "https://view.awsapps.com/start"
-	
+
+	// Default region for IDC
+	defaultIDCRegion = "us-east-1"
+
 	// Polling interval
 	pollInterval = 5 * time.Second
-	
+
 	// Authorization code flow callback
 	authCodeCallbackPath = "/oauth/callback"
 	authCodeCallbackPort = 19877
-	
+
 	// User-Agent to match official Kiro IDE
 	kiroUserAgent = "KiroIDE"
+
+	// IDC token refresh headers (matching Kiro IDE behavior)
+	idcAmzUserAgent = "aws-sdk-js/3.738.0 ua/2.1 os/other lang/js md/browser#unknown_unknown api/sso-oidc#3.738.0 m/E KiroIDE"
+)
+
+// Sentinel errors for OIDC token polling
+var (
+	ErrAuthorizationPending = errors.New("authorization_pending")
+	ErrSlowDown             = errors.New("slow_down")
 )
 
 // SSOOIDCClient handles AWS SSO OIDC authentication.
@@ -83,6 +98,440 @@ type CreateTokenResponse struct {
 	RefreshToken string `json:"refreshToken"`
 }
 
+// getOIDCEndpoint returns the OIDC endpoint for the given region.
+func getOIDCEndpoint(region string) string {
+	if region == "" {
+		region = defaultIDCRegion
+	}
+	return fmt.Sprintf("https://oidc.%s.amazonaws.com", region)
+}
+
+// promptInput prompts the user for input with an optional default value.
+func promptInput(prompt, defaultValue string) string {
+	reader := bufio.NewReader(os.Stdin)
+	if defaultValue != "" {
+		fmt.Printf("%s [%s]: ", prompt, defaultValue)
+	} else {
+		fmt.Printf("%s: ", prompt)
+	}
+	input, err := reader.ReadString('\n')
+	if err != nil {
+		log.Warnf("Error reading input: %v", err)
+		return defaultValue
+	}
+	input = strings.TrimSpace(input)
+	if input == "" {
+		return defaultValue
+	}
+	return input
+}
+
+// promptSelect prompts the user to select from options using number input.
+func promptSelect(prompt string, options []string) int {
+	reader := bufio.NewReader(os.Stdin)
+
+	for {
+		fmt.Println(prompt)
+		for i, opt := range options {
+			fmt.Printf("  %d) %s\n", i+1, opt)
+		}
+		fmt.Printf("Enter selection (1-%d): ", len(options))
+
+		input, err := reader.ReadString('\n')
+		if err != nil {
+			log.Warnf("Error reading input: %v", err)
+			return 0 // Default to first option on error
+		}
+		input = strings.TrimSpace(input)
+
+		// Parse the selection
+		var selection int
+		if _, err := fmt.Sscanf(input, "%d", &selection); err != nil || selection < 1 || selection > len(options) {
+			fmt.Printf("Invalid selection '%s'. Please enter a number between 1 and %d.\n\n", input, len(options))
+			continue
+		}
+		return selection - 1
+	}
+}
+
+// RegisterClientWithRegion registers a new OIDC client with AWS using a specific region.
+func (c *SSOOIDCClient) RegisterClientWithRegion(ctx context.Context, region string) (*RegisterClientResponse, error) {
+	endpoint := getOIDCEndpoint(region)
+
+	payload := map[string]interface{}{
+		"clientName": "Kiro IDE",
+		"clientType": "public",
+		"scopes":     []string{"codewhisperer:completions", "codewhisperer:analysis", "codewhisperer:conversations", "codewhisperer:transformations", "codewhisperer:taskassist"},
+		"grantTypes": []string{"urn:ietf:params:oauth:grant-type:device_code", "refresh_token"},
+	}
+
+	body, err := json.Marshal(payload)
+	if err != nil {
+		return nil, err
+	}
+
+	req, err := http.NewRequestWithContext(ctx, http.MethodPost, endpoint+"/client/register", strings.NewReader(string(body)))
+	if err != nil {
+		return nil, err
+	}
+	req.Header.Set("Content-Type", "application/json")
+	req.Header.Set("User-Agent", kiroUserAgent)
+
+	resp, err := c.httpClient.Do(req)
+	if err != nil {
+		return nil, err
+	}
+	defer resp.Body.Close()
+
+	respBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return nil, err
+	}
+
+	if resp.StatusCode != http.StatusOK {
+		log.Debugf("register client failed (status %d): %s", resp.StatusCode, string(respBody))
+		return nil, fmt.Errorf("register client failed (status %d)", resp.StatusCode)
+	}
+
+	var result RegisterClientResponse
+	if err := json.Unmarshal(respBody, &result); err != nil {
+		return nil, err
+	}
+
+	return &result, nil
+}
+
+// StartDeviceAuthorizationWithIDC starts the device authorization flow for IDC.
+func (c *SSOOIDCClient) StartDeviceAuthorizationWithIDC(ctx context.Context, clientID, clientSecret, startURL, region string) (*StartDeviceAuthResponse, error) {
+	endpoint := getOIDCEndpoint(region)
+
+	payload := map[string]string{
+		"clientId":     clientID,
+		"clientSecret": clientSecret,
+		"startUrl":     startURL,
+	}
+
+	body, err := json.Marshal(payload)
+	if err != nil {
+		return nil, err
+	}
+
+	req, err := http.NewRequestWithContext(ctx, http.MethodPost, endpoint+"/device_authorization", strings.NewReader(string(body)))
+	if err != nil {
+		return nil, err
+	}
+	req.Header.Set("Content-Type", "application/json")
+	req.Header.Set("User-Agent", kiroUserAgent)
+
+	resp, err := c.httpClient.Do(req)
+	if err != nil {
+		return nil, err
+	}
+	defer resp.Body.Close()
+
+	respBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return nil, err
+	}
+
+	if resp.StatusCode != http.StatusOK {
+		log.Debugf("start device auth failed (status %d): %s", resp.StatusCode, string(respBody))
+		return nil, fmt.Errorf("start device auth failed (status %d)", resp.StatusCode)
+	}
+
+	var result StartDeviceAuthResponse
+	if err := json.Unmarshal(respBody, &result); err != nil {
+		return nil, err
+	}
+
+	return &result, nil
+}
+
+// CreateTokenWithRegion polls for the access token after user authorization using a specific region.
+func (c *SSOOIDCClient) CreateTokenWithRegion(ctx context.Context, clientID, clientSecret, deviceCode, region string) (*CreateTokenResponse, error) {
+	endpoint := getOIDCEndpoint(region)
+
+	payload := map[string]string{
+		"clientId":     clientID,
+		"clientSecret": clientSecret,
+		"deviceCode":   deviceCode,
+		"grantType":    "urn:ietf:params:oauth:grant-type:device_code",
+	}
+
+	body, err := json.Marshal(payload)
+	if err != nil {
+		return nil, err
+	}
+
+	req, err := http.NewRequestWithContext(ctx, http.MethodPost, endpoint+"/token", strings.NewReader(string(body)))
+	if err != nil {
+		return nil, err
+	}
+	req.Header.Set("Content-Type", "application/json")
+	req.Header.Set("User-Agent", kiroUserAgent)
+
+	resp, err := c.httpClient.Do(req)
+	if err != nil {
+		return nil, err
+	}
+	defer resp.Body.Close()
+
+	respBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return nil, err
+	}
+
+	// Check for pending authorization
+	if resp.StatusCode == http.StatusBadRequest {
+		var errResp struct {
+			Error string `json:"error"`
+		}
+		if json.Unmarshal(respBody, &errResp) == nil {
+			if errResp.Error == "authorization_pending" {
+				return nil, ErrAuthorizationPending
+			}
+			if errResp.Error == "slow_down" {
+				return nil, ErrSlowDown
+			}
+		}
+		log.Debugf("create token failed: %s", string(respBody))
+		return nil, fmt.Errorf("create token failed")
+	}
+
+	if resp.StatusCode != http.StatusOK {
+		log.Debugf("create token failed (status %d): %s", resp.StatusCode, string(respBody))
+		return nil, fmt.Errorf("create token failed (status %d)", resp.StatusCode)
+	}
+
+	var result CreateTokenResponse
+	if err := json.Unmarshal(respBody, &result); err != nil {
+		return nil, err
+	}
+
+	return &result, nil
+}
+
+// RefreshTokenWithRegion refreshes an access token using the refresh token with a specific region.
+func (c *SSOOIDCClient) RefreshTokenWithRegion(ctx context.Context, clientID, clientSecret, refreshToken, region, startURL string) (*KiroTokenData, error) {
+	endpoint := getOIDCEndpoint(region)
+
+	payload := map[string]string{
+		"clientId":     clientID,
+		"clientSecret": clientSecret,
+		"refreshToken": refreshToken,
+		"grantType":    "refresh_token",
+	}
+
+	body, err := json.Marshal(payload)
+	if err != nil {
+		return nil, err
+	}
+
+	req, err := http.NewRequestWithContext(ctx, http.MethodPost, endpoint+"/token", strings.NewReader(string(body)))
+	if err != nil {
+		return nil, err
+	}
+
+	// Set headers matching kiro2api's IDC token refresh
+	// These headers are required for successful IDC token refresh
+	req.Header.Set("Content-Type", "application/json")
+	req.Header.Set("Host", fmt.Sprintf("oidc.%s.amazonaws.com", region))
+	req.Header.Set("Connection", "keep-alive")
+	req.Header.Set("x-amz-user-agent", idcAmzUserAgent)
+	req.Header.Set("Accept", "*/*")
+	req.Header.Set("Accept-Language", "*")
+	req.Header.Set("sec-fetch-mode", "cors")
+	req.Header.Set("User-Agent", "node")
+	req.Header.Set("Accept-Encoding", "br, gzip, deflate")
+
+	resp, err := c.httpClient.Do(req)
+	if err != nil {
+		return nil, err
+	}
+	defer resp.Body.Close()
+
+	respBody, err := io.ReadAll(resp.Body)
+	if err != nil {
+		return nil, err
+	}
+
+	if resp.StatusCode != http.StatusOK {
+		log.Warnf("IDC token refresh failed (status %d): %s", resp.StatusCode, string(respBody))
+		return nil, fmt.Errorf("token refresh failed (status %d)", resp.StatusCode)
+	}
+
+	var result CreateTokenResponse
+	if err := json.Unmarshal(respBody, &result); err != nil {
+		return nil, err
+	}
+
+	expiresAt := time.Now().Add(time.Duration(result.ExpiresIn) * time.Second)
+
+	return &KiroTokenData{
+		AccessToken:  result.AccessToken,
+		RefreshToken: result.RefreshToken,
+		ExpiresAt:    expiresAt.Format(time.RFC3339),
+		AuthMethod:   "idc",
+		Provider:     "AWS",
+		ClientID:     clientID,
+		ClientSecret: clientSecret,
+		StartURL:     startURL,
+		Region:       region,
+	}, nil
+}
+
+// LoginWithIDC performs the full device code flow for AWS Identity Center (IDC).
+func (c *SSOOIDCClient) LoginWithIDC(ctx context.Context, startURL, region string) (*KiroTokenData, error) {
+	fmt.Println("\n╔══════════════════════════════════════════════════════════╗")
+	fmt.Println("║       Kiro Authentication (AWS Identity Center)          ║")
+	fmt.Println("╚══════════════════════════════════════════════════════════╝")
+
+	// Step 1: Register client with the specified region
+	fmt.Println("\nRegistering client...")
+	regResp, err := c.RegisterClientWithRegion(ctx, region)
+	if err != nil {
+		return nil, fmt.Errorf("failed to register client: %w", err)
+	}
+	log.Debugf("Client registered: %s", regResp.ClientID)
+
+	// Step 2: Start device authorization with IDC start URL
+	fmt.Println("Starting device authorization...")
+	authResp, err := c.StartDeviceAuthorizationWithIDC(ctx, regResp.ClientID, regResp.ClientSecret, startURL, region)
+	if err != nil {
+		return nil, fmt.Errorf("failed to start device auth: %w", err)
+	}
+
+	// Step 3: Show user the verification URL
+	fmt.Printf("\n")
+	fmt.Println("════════════════════════════════════════════════════════════")
+	fmt.Printf("  Confirm the following code in the browser:\n")
+	fmt.Printf("  Code: %s\n", authResp.UserCode)
+	fmt.Println("════════════════════════════════════════════════════════════")
+	fmt.Printf("\n  Open this URL: %s\n\n", authResp.VerificationURIComplete)
+
+	// Set incognito mode based on config
+	if c.cfg != nil {
+		browser.SetIncognitoMode(c.cfg.IncognitoBrowser)
+		if !c.cfg.IncognitoBrowser {
+			log.Info("kiro: using normal browser mode (--no-incognito). Note: You may not be able to select a different account.")
+		} else {
+			log.Debug("kiro: using incognito mode for multi-account support")
+		}
+	} else {
+		browser.SetIncognitoMode(true)
+		log.Debug("kiro: using incognito mode for multi-account support (default)")
+	}
+
+	// Open browser
+	if err := browser.OpenURL(authResp.VerificationURIComplete); err != nil {
+		log.Warnf("Could not open browser automatically: %v", err)
+		fmt.Println("  Please open the URL manually in your browser.")
+	} else {
+		fmt.Println("  (Browser opened automatically)")
+	}
+
+	// Step 4: Poll for token
+	fmt.Println("Waiting for authorization...")
+
+	interval := pollInterval
+	if authResp.Interval > 0 {
+		interval = time.Duration(authResp.Interval) * time.Second
+	}
+
+	deadline := time.Now().Add(time.Duration(authResp.ExpiresIn) * time.Second)
+
+	for time.Now().Before(deadline) {
+		select {
+		case <-ctx.Done():
+			browser.CloseBrowser()
+			return nil, ctx.Err()
+		case <-time.After(interval):
+			tokenResp, err := c.CreateTokenWithRegion(ctx, regResp.ClientID, regResp.ClientSecret, authResp.DeviceCode, region)
+			if err != nil {
+				if errors.Is(err, ErrAuthorizationPending) {
+					fmt.Print(".")
+					continue
+				}
+				if errors.Is(err, ErrSlowDown) {
+					interval += 5 * time.Second
+					continue
+				}
+				browser.CloseBrowser()
+				return nil, fmt.Errorf("token creation failed: %w", err)
+			}
+
+			fmt.Println("\n\n✓ Authorization successful!")
+
+			// Close the browser window
+			if err := browser.CloseBrowser(); err != nil {
+				log.Debugf("Failed to close browser: %v", err)
+			}
+
+			// Step 5: Get profile ARN from CodeWhisperer API
+			fmt.Println("Fetching profile information...")
+			profileArn := c.fetchProfileArn(ctx, tokenResp.AccessToken)
+
+			// Fetch user email
+			email := FetchUserEmailWithFallback(ctx, c.cfg, tokenResp.AccessToken)
+			if email != "" {
+				fmt.Printf("  Logged in as: %s\n", email)
+			}
+
+			expiresAt := time.Now().Add(time.Duration(tokenResp.ExpiresIn) * time.Second)
+
+			return &KiroTokenData{
+				AccessToken:  tokenResp.AccessToken,
+				RefreshToken: tokenResp.RefreshToken,
+				ProfileArn:   profileArn,
+				ExpiresAt:    expiresAt.Format(time.RFC3339),
+				AuthMethod:   "idc",
+				Provider:     "AWS",
+				ClientID:     regResp.ClientID,
+				ClientSecret: regResp.ClientSecret,
+				Email:        email,
+				StartURL:     startURL,
+				Region:       region,
+			}, nil
+		}
+	}
+
+	// Close browser on timeout
+	if err := browser.CloseBrowser(); err != nil {
+		log.Debugf("Failed to close browser on timeout: %v", err)
+	}
+	return nil, fmt.Errorf("authorization timed out")
+}
+
+// LoginWithMethodSelection prompts the user to select between Builder ID and IDC, then performs the login.
+func (c *SSOOIDCClient) LoginWithMethodSelection(ctx context.Context) (*KiroTokenData, error) {
+	fmt.Println("\n╔══════════════════════════════════════════════════════════╗")
+	fmt.Println("║              Kiro Authentication (AWS)                    ║")
+	fmt.Println("╚══════════════════════════════════════════════════════════╝")
+
+	// Prompt for login method
+	options := []string{
+		"Use with Builder ID (personal AWS account)",
+		"Use with IDC Account (organization SSO)",
+	}
+	selection := promptSelect("\n? Select login method:", options)
+
+	if selection == 0 {
+		// Builder ID flow - use existing implementation
+		return c.LoginWithBuilderID(ctx)
+	}
+
+	// IDC flow - prompt for start URL and region
+	fmt.Println()
+	startURL := promptInput("? Enter Start URL", "")
+	if startURL == "" {
+		return nil, fmt.Errorf("start URL is required for IDC login")
+	}
+
+	region := promptInput("? Enter Region", defaultIDCRegion)
+
+	return c.LoginWithIDC(ctx, startURL, region)
+}
+
 // RegisterClient registers a new OIDC client with AWS.
 func (c *SSOOIDCClient) RegisterClient(ctx context.Context) (*RegisterClientResponse, error) {
 	payload := map[string]interface{}{
@@ -211,10 +660,10 @@ func (c *SSOOIDCClient) CreateToken(ctx context.Context, clientID, clientSecret,
 		}
 		if json.Unmarshal(respBody, &errResp) == nil {
 			if errResp.Error == "authorization_pending" {
-				return nil, fmt.Errorf("authorization_pending")
+				return nil, ErrAuthorizationPending
 			}
 			if errResp.Error == "slow_down" {
-				return nil, fmt.Errorf("slow_down")
+				return nil, ErrSlowDown
 			}
 		}
 		log.Debugf("create token failed: %s", string(respBody))
@@ -359,12 +808,11 @@ func (c *SSOOIDCClient) LoginWithBuilderID(ctx context.Context) (*KiroTokenData,
 		case <-time.After(interval):
 			tokenResp, err := c.CreateToken(ctx, regResp.ClientID, regResp.ClientSecret, authResp.DeviceCode)
 			if err != nil {
-				errStr := err.Error()
-				if strings.Contains(errStr, "authorization_pending") {
+				if errors.Is(err, ErrAuthorizationPending) {
 					fmt.Print(".")
 					continue
 				}
-				if strings.Contains(errStr, "slow_down") {
+				if errors.Is(err, ErrSlowDown) {
 					interval += 5 * time.Second
 					continue
 				}

internal/config/config.go

@@ -39,6 +39,9 @@ type Config struct {
 	// Debug enables or disables debug-level logging and other debug features.
 	Debug bool `yaml:"debug" json:"debug"`
 
+	// CommercialMode disables high-overhead HTTP middleware features to minimize per-request memory usage.
+	CommercialMode bool `yaml:"commercial-mode" json:"commercial-mode"`
+
 	// LoggingToFile controls whether application logs are written to rotating files or stdout.
 	LoggingToFile bool `yaml:"logging-to-file" json:"logging-to-file"`
 
@@ -60,6 +63,9 @@ type Config struct {
 	// QuotaExceeded defines the behavior when a quota is exceeded.
 	QuotaExceeded QuotaExceeded `yaml:"quota-exceeded" json:"quota-exceeded"`
 
+	// Routing controls credential selection behavior.
+	Routing RoutingConfig `yaml:"routing" json:"routing"`
+
 	// WebsocketAuth enables or disables authentication for the WebSocket API.
 	WebsocketAuth bool `yaml:"ws-auth" json:"ws-auth"`
 
@@ -92,6 +98,14 @@ type Config struct {
 	// OAuthExcludedModels defines per-provider global model exclusions applied to OAuth/file-backed auth entries.
 	OAuthExcludedModels map[string][]string `yaml:"oauth-excluded-models,omitempty" json:"oauth-excluded-models,omitempty"`
 
+	// OAuthModelMappings defines global model name mappings for OAuth/file-backed auth channels.
+	// These mappings affect both model listing and model routing for supported channels:
+	// gemini-cli, vertex, aistudio, antigravity, claude, codex, qwen, iflow.
+	//
+	// NOTE: This does not apply to existing per-credential model alias features under:
+	// gemini-api-key, codex-api-key, claude-api-key, openai-compatibility, vertex-api-key, and ampcode.
+	OAuthModelMappings map[string][]ModelNameMapping `yaml:"oauth-model-mappings,omitempty" json:"oauth-model-mappings,omitempty"`
+
 	// Payload defines default and override rules for provider payload parameters.
 	Payload PayloadConfig `yaml:"payload" json:"payload"`
 
@@ -136,6 +150,20 @@ type QuotaExceeded struct {
 	SwitchPreviewModel bool `yaml:"switch-preview-model" json:"switch-preview-model"`
 }
 
+// RoutingConfig configures how credentials are selected for requests.
+type RoutingConfig struct {
+	// Strategy selects the credential selection strategy.
+	// Supported values: "round-robin" (default), "fill-first".
+	Strategy string `yaml:"strategy,omitempty" json:"strategy,omitempty"`
+}
+
+// ModelNameMapping defines a model ID rename mapping for a specific channel.
+// It maps the original model name (Name) to the client-visible alias (Alias).
+type ModelNameMapping struct {
+	Name  string `yaml:"name" json:"name"`
+	Alias string `yaml:"alias" json:"alias"`
+}
+
 // AmpModelMapping defines a model name mapping for Amp CLI requests.
 // When Amp requests a model that isn't available locally, this mapping
 // allows routing to an alternative model that IS available.
@@ -146,6 +174,11 @@ type AmpModelMapping struct {
 	// To is the target model name to route to (e.g., "claude-sonnet-4").
 	// The target model must have available providers in the registry.
 	To string `yaml:"to" json:"to"`
+
+	// Regex indicates whether the 'from' field should be interpreted as a regular
+	// expression for matching model names. When true, this mapping is evaluated
+	// after exact matches and in the order provided. Defaults to false (exact match).
+	Regex bool `yaml:"regex,omitempty" json:"regex,omitempty"`
 }
 
 // AmpCode groups Amp CLI integration settings including upstream routing,
@@ -157,6 +190,11 @@ type AmpCode struct {
 	// UpstreamAPIKey optionally overrides the Authorization header when proxying Amp upstream calls.
 	UpstreamAPIKey string `yaml:"upstream-api-key" json:"upstream-api-key"`
 
+	// UpstreamAPIKeys maps client API keys (from top-level api-keys) to upstream API keys.
+	// When a client authenticates with a key that matches an entry, that upstream key is used.
+	// If no match is found, falls back to UpstreamAPIKey (default behavior).
+	UpstreamAPIKeys []AmpUpstreamAPIKeyEntry `yaml:"upstream-api-keys,omitempty" json:"upstream-api-keys,omitempty"`
+
 	// RestrictManagementToLocalhost restricts Amp management routes (/api/user, /api/threads, etc.)
 	// to only accept connections from localhost (127.0.0.1, ::1). When true, prevents drive-by
 	// browser attacks and remote access to management endpoints. Default: false (API key auth is sufficient).
@@ -172,6 +210,17 @@ type AmpCode struct {
 	ForceModelMappings bool `yaml:"force-model-mappings" json:"force-model-mappings"`
 }
 
+// AmpUpstreamAPIKeyEntry maps a set of client API keys to a specific upstream API key.
+// When a request is authenticated with one of the APIKeys, the corresponding UpstreamAPIKey
+// is used for the upstream Amp request.
+type AmpUpstreamAPIKeyEntry struct {
+	// UpstreamAPIKey is the API key to use when proxying to the Amp upstream.
+	UpstreamAPIKey string `yaml:"upstream-api-key" json:"upstream-api-key"`
+
+	// APIKeys are the client API keys (from top-level api-keys) that map to this upstream key.
+	APIKeys []string `yaml:"api-keys" json:"api-keys"`
+}
+
 // PayloadConfig defines default and override parameter rules applied to provider payloads.
 type PayloadConfig struct {
 	// Default defines rules that only set parameters when they are missing in the payload.
@@ -231,6 +280,9 @@ type ClaudeModel struct {
 	Alias string `yaml:"alias" json:"alias"`
 }
 
+func (m ClaudeModel) GetName() string  { return m.Name }
+func (m ClaudeModel) GetAlias() string { return m.Alias }
+
 // CodexKey represents the configuration for a Codex API key,
 // including the API key itself and an optional base URL for the API endpoint.
 type CodexKey struct {
@@ -247,6 +299,9 @@ type CodexKey struct {
 	// ProxyURL overrides the global proxy setting for this API key if provided.
 	ProxyURL string `yaml:"proxy-url" json:"proxy-url"`
 
+	// Models defines upstream model names and aliases for request routing.
+	Models []CodexModel `yaml:"models" json:"models"`
+
 	// Headers optionally adds extra HTTP headers for requests sent with this key.
 	Headers map[string]string `yaml:"headers,omitempty" json:"headers,omitempty"`
 
@@ -254,6 +309,18 @@ type CodexKey struct {
 	ExcludedModels []string `yaml:"excluded-models,omitempty" json:"excluded-models,omitempty"`
 }
 
+// CodexModel describes a mapping between an alias and the actual upstream model name.
+type CodexModel struct {
+	// Name is the upstream model identifier used when issuing requests.
+	Name string `yaml:"name" json:"name"`
+
+	// Alias is the client-facing model name that maps to Name.
+	Alias string `yaml:"alias" json:"alias"`
+}
+
+func (m CodexModel) GetName() string  { return m.Name }
+func (m CodexModel) GetAlias() string { return m.Alias }
+
 // GeminiKey represents the configuration for a Gemini API key,
 // including optional overrides for upstream base URL, proxy routing, and headers.
 type GeminiKey struct {
@@ -269,6 +336,9 @@ type GeminiKey struct {
 	// ProxyURL optionally overrides the global proxy for this API key.
 	ProxyURL string `yaml:"proxy-url,omitempty" json:"proxy-url,omitempty"`
 
+	// Models defines upstream model names and aliases for request routing.
+	Models []GeminiModel `yaml:"models,omitempty" json:"models,omitempty"`
+
 	// Headers optionally adds extra HTTP headers for requests sent with this key.
 	Headers map[string]string `yaml:"headers,omitempty" json:"headers,omitempty"`
 
@@ -276,6 +346,18 @@ type GeminiKey struct {
 	ExcludedModels []string `yaml:"excluded-models,omitempty" json:"excluded-models,omitempty"`
 }
 
+// GeminiModel describes a mapping between an alias and the actual upstream model name.
+type GeminiModel struct {
+	// Name is the upstream model identifier used when issuing requests.
+	Name string `yaml:"name" json:"name"`
+
+	// Alias is the client-facing model name that maps to Name.
+	Alias string `yaml:"alias" json:"alias"`
+}
+
+func (m GeminiModel) GetName() string  { return m.Name }
+func (m GeminiModel) GetAlias() string { return m.Alias }
+
 // KiroKey represents the configuration for Kiro (AWS CodeWhisperer) authentication.
 type KiroKey struct {
 	// TokenFile is the path to the Kiro token file (default: ~/.aws/sso/cache/kiro-auth-token.json)
@@ -391,7 +473,7 @@ func LoadConfigOptional(configFile string, optional bool) (*Config, error) {
 	cfg.DisableCooling = false
 	cfg.AmpCode.RestrictManagementToLocalhost = false // Default to false: API key auth is sufficient
 	cfg.RemoteManagement.PanelGitHubRepository = DefaultPanelGitHubRepository
-	cfg.IncognitoBrowser = false                     // Default to normal browser (AWS uses incognito by force)
+	cfg.IncognitoBrowser = false // Default to normal browser (AWS uses incognito by force)
 	if err = yaml.Unmarshal(data, &cfg); err != nil {
 		if optional {
 			// In cloud deploy mode, if YAML parsing fails, return empty config instead of error.
@@ -460,6 +542,9 @@ func LoadConfigOptional(configFile string, optional bool) (*Config, error) {
 	// Normalize OAuth provider model exclusion map.
 	cfg.OAuthExcludedModels = NormalizeOAuthExcludedModels(cfg.OAuthExcludedModels)
 
+	// Normalize global OAuth model name mappings.
+	cfg.SanitizeOAuthModelMappings()
+
 	if cfg.legacyMigrationPending {
 		fmt.Println("Detected legacy configuration keys, attempting to persist the normalized config...")
 		if !optional && configFile != "" {
@@ -476,6 +561,50 @@ func LoadConfigOptional(configFile string, optional bool) (*Config, error) {
 	return &cfg, nil
 }
 
+// SanitizeOAuthModelMappings normalizes and deduplicates global OAuth model name mappings.
+// It trims whitespace, normalizes channel keys to lower-case, drops empty entries,
+// and ensures (From, To) pairs are unique within each channel.
+func (cfg *Config) SanitizeOAuthModelMappings() {
+	if cfg == nil || len(cfg.OAuthModelMappings) == 0 {
+		return
+	}
+	out := make(map[string][]ModelNameMapping, len(cfg.OAuthModelMappings))
+	for rawChannel, mappings := range cfg.OAuthModelMappings {
+		channel := strings.ToLower(strings.TrimSpace(rawChannel))
+		if channel == "" || len(mappings) == 0 {
+			continue
+		}
+		seenName := make(map[string]struct{}, len(mappings))
+		seenAlias := make(map[string]struct{}, len(mappings))
+		clean := make([]ModelNameMapping, 0, len(mappings))
+		for _, mapping := range mappings {
+			name := strings.TrimSpace(mapping.Name)
+			alias := strings.TrimSpace(mapping.Alias)
+			if name == "" || alias == "" {
+				continue
+			}
+			if strings.EqualFold(name, alias) {
+				continue
+			}
+			nameKey := strings.ToLower(name)
+			aliasKey := strings.ToLower(alias)
+			if _, ok := seenName[nameKey]; ok {
+				continue
+			}
+			if _, ok := seenAlias[aliasKey]; ok {
+				continue
+			}
+			seenName[nameKey] = struct{}{}
+			seenAlias[aliasKey] = struct{}{}
+			clean = append(clean, ModelNameMapping{Name: name, Alias: alias})
+		}
+		if len(clean) > 0 {
+			out[channel] = clean
+		}
+	}
+	cfg.OAuthModelMappings = out
+}
+
 // SanitizeOpenAICompatibility removes OpenAI-compatibility provider entries that are
 // not actionable, specifically those missing a BaseURL. It trims whitespace before
 // evaluation and preserves the relative order of remaining entries.
@@ -861,8 +990,8 @@ func getOrCreateMapValue(mapNode *yaml.Node, key string) *yaml.Node {
 }
 
 // mergeMappingPreserve merges keys from src into dst mapping node while preserving
-// key order and comments of existing keys in dst. Unknown keys from src are appended
-// to dst at the end, copying their node structure from src.
+// key order and comments of existing keys in dst. New keys are only added if their
+// value is non-zero to avoid polluting the config with defaults.
 func mergeMappingPreserve(dst, src *yaml.Node) {
 	if dst == nil || src == nil {
 		return
@@ -873,20 +1002,19 @@ func mergeMappingPreserve(dst, src *yaml.Node) {
 		copyNodeShallow(dst, src)
 		return
 	}
-	// Build a lookup of existing keys in dst
 	for i := 0; i+1 < len(src.Content); i += 2 {
 		sk := src.Content[i]
 		sv := src.Content[i+1]
 		idx := findMapKeyIndex(dst, sk.Value)
 		if idx >= 0 {
-			// Merge into existing value node
+			// Merge into existing value node (always update, even to zero values)
 			dv := dst.Content[idx+1]
 			mergeNodePreserve(dv, sv)
 		} else {
-			if shouldSkipEmptyCollectionOnPersist(sk.Value, sv) {
+			// New key: only add if value is non-zero to avoid polluting config with defaults
+			if isZeroValueNode(sv) {
 				continue
 			}
-			// Append new key/value pair by deep-copying from src
 			dst.Content = append(dst.Content, deepCopyNode(sk), deepCopyNode(sv))
 		}
 	}
@@ -969,32 +1097,49 @@ func findMapKeyIndex(mapNode *yaml.Node, key string) int {
 	return -1
 }
 
-func shouldSkipEmptyCollectionOnPersist(key string, node *yaml.Node) bool {
-	switch key {
-	case "generative-language-api-key",
-		"gemini-api-key",
-		"vertex-api-key",
-		"claude-api-key",
-		"codex-api-key",
-		"openai-compatibility":
-		return isEmptyCollectionNode(node)
-	default:
-		return false
-	}
-}
-
-func isEmptyCollectionNode(node *yaml.Node) bool {
+// isZeroValueNode returns true if the YAML node represents a zero/default value
+// that should not be written as a new key to preserve config cleanliness.
+// For mappings and sequences, recursively checks if all children are zero values.
+func isZeroValueNode(node *yaml.Node) bool {
 	if node == nil {
 		return true
 	}
 	switch node.Kind {
-	case yaml.SequenceNode:
-		return len(node.Content) == 0
 	case yaml.ScalarNode:
-		return node.Tag == "!!null"
-	default:
-		return false
+		switch node.Tag {
+		case "!!bool":
+			return node.Value == "false"
+		case "!!int", "!!float":
+			return node.Value == "0" || node.Value == "0.0"
+		case "!!str":
+			return node.Value == ""
+		case "!!null":
+			return true
+		}
+	case yaml.SequenceNode:
+		if len(node.Content) == 0 {
+			return true
+		}
+		// Check if all elements are zero values
+		for _, child := range node.Content {
+			if !isZeroValueNode(child) {
+				return false
+			}
+		}
+		return true
+	case yaml.MappingNode:
+		if len(node.Content) == 0 {
+			return true
+		}
+		// Check if all values are zero values (values are at odd indices)
+		for i := 1; i < len(node.Content); i += 2 {
+			if !isZeroValueNode(node.Content[i]) {
+				return false
+			}
+		}
+		return true
 	}
+	return false
 }
 
 // deepCopyNode creates a deep copy of a yaml.Node graph.

internal/config/sdk_config.go

@@ -22,6 +22,21 @@ type SDKConfig struct {
 
 	// Access holds request authentication provider configuration.
 	Access AccessConfig `yaml:"auth,omitempty" json:"auth,omitempty"`
+
+	// Streaming configures server-side streaming behavior (keep-alives and safe bootstrap retries).
+	Streaming StreamingConfig `yaml:"streaming" json:"streaming"`
+}
+
+// StreamingConfig holds server streaming behavior configuration.
+type StreamingConfig struct {
+	// KeepAliveSeconds controls how often the server emits SSE heartbeats (": keep-alive\n\n").
+	// <= 0 disables keep-alives. Default is 0.
+	KeepAliveSeconds int `yaml:"keepalive-seconds,omitempty" json:"keepalive-seconds,omitempty"`
+
+	// BootstrapRetries controls how many times the server may retry a streaming request before any bytes are sent,
+	// to allow auth rotation / transient recovery.
+	// <= 0 disables bootstrap retries. Default is 0.
+	BootstrapRetries int `yaml:"bootstrap-retries,omitempty" json:"bootstrap-retries,omitempty"`
 }
 
 // AccessConfig groups request authentication providers.

internal/config/vertex_compat.go

@@ -42,6 +42,9 @@ type VertexCompatModel struct {
 	Alias string `yaml:"alias" json:"alias"`
 }
 
+func (m VertexCompatModel) GetName() string  { return m.Name }
+func (m VertexCompatModel) GetAlias() string { return m.Alias }
+
 // SanitizeVertexCompatKeys deduplicates and normalizes Vertex-compatible API key credentials.
 func (cfg *Config) SanitizeVertexCompatKeys() {
 	if cfg == nil {

internal/logging/gin_logger.go

@@ -7,6 +7,7 @@ import (
 	"fmt"
 	"net/http"
 	"runtime/debug"
+	"strings"
 	"time"
 
 	"github.com/gin-gonic/gin"
@@ -14,11 +15,24 @@ import (
 	log "github.com/sirupsen/logrus"
 )
 
+// aiAPIPrefixes defines path prefixes for AI API requests that should have request ID tracking.
+var aiAPIPrefixes = []string{
+	"/v1/chat/completions",
+	"/v1/completions",
+	"/v1/messages",
+	"/v1/responses",
+	"/v1beta/models/",
+	"/api/provider/",
+}
+
 const skipGinLogKey = "__gin_skip_request_logging__"
 
 // GinLogrusLogger returns a Gin middleware handler that logs HTTP requests and responses
 // using logrus. It captures request details including method, path, status code, latency,
-// client IP, and any error messages, formatting them in a Gin-style log format.
+// client IP, and any error messages. Request ID is only added for AI API requests.
+//
+// Output format (AI API): [2025-12-23 20:14:10] [info ] | a1b2c3d4 | 200 |       23.559s | ...
+// Output format (others): [2025-12-23 20:14:10] [info ] | -------- | 200 |       23.559s | ...
 //
 // Returns:
 //   - gin.HandlerFunc: A middleware handler for request logging
@@ -28,6 +42,15 @@ func GinLogrusLogger() gin.HandlerFunc {
 		path := c.Request.URL.Path
 		raw := util.MaskSensitiveQuery(c.Request.URL.RawQuery)
 
+		// Only generate request ID for AI API paths
+		var requestID string
+		if isAIAPIPath(path) {
+			requestID = GenerateRequestID()
+			SetGinRequestID(c, requestID)
+			ctx := WithRequestID(c.Request.Context(), requestID)
+			c.Request = c.Request.WithContext(ctx)
+		}
+
 		c.Next()
 
 		if shouldSkipGinRequestLogging(c) {
@@ -49,23 +72,38 @@ func GinLogrusLogger() gin.HandlerFunc {
 		clientIP := c.ClientIP()
 		method := c.Request.Method
 		errorMessage := c.Errors.ByType(gin.ErrorTypePrivate).String()
-		timestamp := time.Now().Format("2006/01/02 - 15:04:05")
-		logLine := fmt.Sprintf("[GIN] %s | %3d | %13v | %15s | %-7s \"%s\"", timestamp, statusCode, latency, clientIP, method, path)
+
+		if requestID == "" {
+			requestID = "--------"
+		}
+		logLine := fmt.Sprintf("%3d | %13v | %15s | %-7s \"%s\"", statusCode, latency, clientIP, method, path)
 		if errorMessage != "" {
 			logLine = logLine + " | " + errorMessage
 		}
 
+		entry := log.WithField("request_id", requestID)
+
 		switch {
 		case statusCode >= http.StatusInternalServerError:
-			log.Error(logLine)
+			entry.Error(logLine)
 		case statusCode >= http.StatusBadRequest:
-			log.Warn(logLine)
+			entry.Warn(logLine)
 		default:
-			log.Info(logLine)
+			entry.Info(logLine)
 		}
 	}
 }
 
+// isAIAPIPath checks if the given path is an AI API endpoint that should have request ID tracking.
+func isAIAPIPath(path string) bool {
+	for _, prefix := range aiAPIPrefixes {
+		if strings.HasPrefix(path, prefix) {
+			return true
+		}
+	}
+	return false
+}
+
 // GinLogrusRecovery returns a Gin middleware handler that recovers from panics and logs
 // them using logrus. When a panic occurs, it captures the panic value, stack trace,
 // and request path, then returns a 500 Internal Server Error response to the client.

internal/logging/global_logger.go

@@ -10,6 +10,7 @@ import (
 	"sync"
 
 	"github.com/gin-gonic/gin"
+	"github.com/router-for-me/CLIProxyAPI/v6/internal/config"
 	"github.com/router-for-me/CLIProxyAPI/v6/internal/util"
 	log "github.com/sirupsen/logrus"
 	"gopkg.in/natefinch/lumberjack.v2"
@@ -24,7 +25,8 @@ var (
 )
 
 // LogFormatter defines a custom log format for logrus.
-// This formatter adds timestamp, level, and source location to each log entry.
+// This formatter adds timestamp, level, request ID, and source location to each log entry.
+// Format: [2025-12-23 20:14:04] [debug] [manager.go:524] | a1b2c3d4 | Use API key sk-9...0RHO for model gpt-5.2
 type LogFormatter struct{}
 
 // Format renders a single log entry with custom formatting.
@@ -38,16 +40,24 @@ func (m *LogFormatter) Format(entry *log.Entry) ([]byte, error) {
 
 	timestamp := entry.Time.Format("2006-01-02 15:04:05")
 	message := strings.TrimRight(entry.Message, "\r\n")
-	
-	// Handle nil Caller (can happen with some log entries)
-	callerFile := "unknown"
-	callerLine := 0
-	if entry.Caller != nil {
-		callerFile = filepath.Base(entry.Caller.File)
-		callerLine = entry.Caller.Line
+
+	reqID := "--------"
+	if id, ok := entry.Data["request_id"].(string); ok && id != "" {
+		reqID = id
+	}
+
+	level := entry.Level.String()
+	if level == "warning" {
+		level = "warn"
+	}
+	levelStr := fmt.Sprintf("%-5s", level)
+
+	var formatted string
+	if entry.Caller != nil {
+		formatted = fmt.Sprintf("[%s] [%s] [%s] [%s:%d] %s\n", timestamp, reqID, levelStr, filepath.Base(entry.Caller.File), entry.Caller.Line, message)
+	} else {
+		formatted = fmt.Sprintf("[%s] [%s] [%s] %s\n", timestamp, reqID, levelStr, message)
 	}
-	
-	formatted := fmt.Sprintf("[%s] [%s] [%s:%d] %s\n", timestamp, entry.Level, callerFile, callerLine, message)
 	buffer.WriteString(formatted)
 
 	return buffer.Bytes(), nil
@@ -75,10 +85,30 @@ func SetupBaseLogger() {
 	})
 }
 
+// isDirWritable checks if the specified directory exists and is writable by attempting to create and remove a test file.
+func isDirWritable(dir string) bool {
+	info, err := os.Stat(dir)
+	if err != nil || !info.IsDir() {
+		return false
+	}
+
+	testFile := filepath.Join(dir, ".perm_test")
+	f, err := os.Create(testFile)
+	if err != nil {
+		return false
+	}
+
+	defer func() {
+		_ = f.Close()
+		_ = os.Remove(testFile)
+	}()
+	return true
+}
+
 // ConfigureLogOutput switches the global log destination between rotating files and stdout.
 // When logsMaxTotalSizeMB > 0, a background cleaner removes the oldest log files in the logs directory
 // until the total size is within the limit.
-func ConfigureLogOutput(loggingToFile bool, logsMaxTotalSizeMB int) error {
+func ConfigureLogOutput(cfg *config.Config) error {
 	SetupBaseLogger()
 
 	writerMu.Lock()
@@ -87,10 +117,12 @@ func ConfigureLogOutput(loggingToFile bool, logsMaxTotalSizeMB int) error {
 	logDir := "logs"
 	if base := util.WritablePath(); base != "" {
 		logDir = filepath.Join(base, "logs")
+	} else if !isDirWritable(logDir) {
+		logDir = filepath.Join(cfg.AuthDir, "logs")
 	}
 
 	protectedPath := ""
-	if loggingToFile {
+	if cfg.LoggingToFile {
 		if err := os.MkdirAll(logDir, 0o755); err != nil {
 			return fmt.Errorf("logging: failed to create log directory: %w", err)
 		}
@@ -114,7 +146,7 @@ func ConfigureLogOutput(loggingToFile bool, logsMaxTotalSizeMB int) error {
 		log.SetOutput(os.Stdout)
 	}
 
-	configureLogDirCleanerLocked(logDir, logsMaxTotalSizeMB, protectedPath)
+	configureLogDirCleanerLocked(logDir, cfg.LogsMaxTotalSizeMB, protectedPath)
 	return nil
 }
 

internal/logging/request_logger.go

@@ -43,10 +43,11 @@ type RequestLogger interface {
 	//   - response: The raw response data
 	//   - apiRequest: The API request data
 	//   - apiResponse: The API response data
+	//   - requestID: Optional request ID for log file naming
 	//
 	// Returns:
 	//   - error: An error if logging fails, nil otherwise
-	LogRequest(url, method string, requestHeaders map[string][]string, body []byte, statusCode int, responseHeaders map[string][]string, response, apiRequest, apiResponse []byte, apiResponseErrors []*interfaces.ErrorMessage) error
+	LogRequest(url, method string, requestHeaders map[string][]string, body []byte, statusCode int, responseHeaders map[string][]string, response, apiRequest, apiResponse []byte, apiResponseErrors []*interfaces.ErrorMessage, requestID string) error
 
 	// LogStreamingRequest initiates logging for a streaming request and returns a writer for chunks.
 	//
@@ -55,11 +56,12 @@ type RequestLogger interface {
 	//   - method: The HTTP method
 	//   - headers: The request headers
 	//   - body: The request body
+	//   - requestID: Optional request ID for log file naming
 	//
 	// Returns:
 	//   - StreamingLogWriter: A writer for streaming response chunks
 	//   - error: An error if logging initialization fails, nil otherwise
-	LogStreamingRequest(url, method string, headers map[string][]string, body []byte) (StreamingLogWriter, error)
+	LogStreamingRequest(url, method string, headers map[string][]string, body []byte, requestID string) (StreamingLogWriter, error)
 
 	// IsEnabled returns whether request logging is currently enabled.
 	//
@@ -177,20 +179,21 @@ func (l *FileRequestLogger) SetEnabled(enabled bool) {
 //   - response: The raw response data
 //   - apiRequest: The API request data
 //   - apiResponse: The API response data
+//   - requestID: Optional request ID for log file naming
 //
 // Returns:
 //   - error: An error if logging fails, nil otherwise
-func (l *FileRequestLogger) LogRequest(url, method string, requestHeaders map[string][]string, body []byte, statusCode int, responseHeaders map[string][]string, response, apiRequest, apiResponse []byte, apiResponseErrors []*interfaces.ErrorMessage) error {
-	return l.logRequest(url, method, requestHeaders, body, statusCode, responseHeaders, response, apiRequest, apiResponse, apiResponseErrors, false)
+func (l *FileRequestLogger) LogRequest(url, method string, requestHeaders map[string][]string, body []byte, statusCode int, responseHeaders map[string][]string, response, apiRequest, apiResponse []byte, apiResponseErrors []*interfaces.ErrorMessage, requestID string) error {
+	return l.logRequest(url, method, requestHeaders, body, statusCode, responseHeaders, response, apiRequest, apiResponse, apiResponseErrors, false, requestID)
 }
 
 // LogRequestWithOptions logs a request with optional forced logging behavior.
 // The force flag allows writing error logs even when regular request logging is disabled.
-func (l *FileRequestLogger) LogRequestWithOptions(url, method string, requestHeaders map[string][]string, body []byte, statusCode int, responseHeaders map[string][]string, response, apiRequest, apiResponse []byte, apiResponseErrors []*interfaces.ErrorMessage, force bool) error {
-	return l.logRequest(url, method, requestHeaders, body, statusCode, responseHeaders, response, apiRequest, apiResponse, apiResponseErrors, force)
+func (l *FileRequestLogger) LogRequestWithOptions(url, method string, requestHeaders map[string][]string, body []byte, statusCode int, responseHeaders map[string][]string, response, apiRequest, apiResponse []byte, apiResponseErrors []*interfaces.ErrorMessage, force bool, requestID string) error {
+	return l.logRequest(url, method, requestHeaders, body, statusCode, responseHeaders, response, apiRequest, apiResponse, apiResponseErrors, force, requestID)
 }
 
-func (l *FileRequestLogger) logRequest(url, method string, requestHeaders map[string][]string, body []byte, statusCode int, responseHeaders map[string][]string, response, apiRequest, apiResponse []byte, apiResponseErrors []*interfaces.ErrorMessage, force bool) error {
+func (l *FileRequestLogger) logRequest(url, method string, requestHeaders map[string][]string, body []byte, statusCode int, responseHeaders map[string][]string, response, apiRequest, apiResponse []byte, apiResponseErrors []*interfaces.ErrorMessage, force bool, requestID string) error {
 	if !l.enabled && !force {
 		return nil
 	}
@@ -200,10 +203,10 @@ func (l *FileRequestLogger) logRequest(url, method string, requestHeaders map[st
 		return fmt.Errorf("failed to create logs directory: %w", errEnsure)
 	}
 
-	// Generate filename
-	filename := l.generateFilename(url)
+	// Generate filename with request ID
+	filename := l.generateFilename(url, requestID)
 	if force && !l.enabled {
-		filename = l.generateErrorFilename(url)
+		filename = l.generateErrorFilename(url, requestID)
 	}
 	filePath := filepath.Join(l.logsDir, filename)
 
@@ -271,11 +274,12 @@ func (l *FileRequestLogger) logRequest(url, method string, requestHeaders map[st
 //   - method: The HTTP method
 //   - headers: The request headers
 //   - body: The request body
+//   - requestID: Optional request ID for log file naming
 //
 // Returns:
 //   - StreamingLogWriter: A writer for streaming response chunks
 //   - error: An error if logging initialization fails, nil otherwise
-func (l *FileRequestLogger) LogStreamingRequest(url, method string, headers map[string][]string, body []byte) (StreamingLogWriter, error) {
+func (l *FileRequestLogger) LogStreamingRequest(url, method string, headers map[string][]string, body []byte, requestID string) (StreamingLogWriter, error) {
 	if !l.enabled {
 		return &NoOpStreamingLogWriter{}, nil
 	}
@@ -285,8 +289,8 @@ func (l *FileRequestLogger) LogStreamingRequest(url, method string, headers map[
 		return nil, fmt.Errorf("failed to create logs directory: %w", err)
 	}
 
-	// Generate filename
-	filename := l.generateFilename(url)
+	// Generate filename with request ID
+	filename := l.generateFilename(url, requestID)
 	filePath := filepath.Join(l.logsDir, filename)
 
 	requestHeaders := make(map[string][]string, len(headers))
@@ -330,8 +334,8 @@ func (l *FileRequestLogger) LogStreamingRequest(url, method string, headers map[
 }
 
 // generateErrorFilename creates a filename with an error prefix to differentiate forced error logs.
-func (l *FileRequestLogger) generateErrorFilename(url string) string {
-	return fmt.Sprintf("error-%s", l.generateFilename(url))
+func (l *FileRequestLogger) generateErrorFilename(url string, requestID ...string) string {
+	return fmt.Sprintf("error-%s", l.generateFilename(url, requestID...))
 }
 
 // ensureLogsDir creates the logs directory if it doesn't exist.
@@ -346,13 +350,15 @@ func (l *FileRequestLogger) ensureLogsDir() error {
 }
 
 // generateFilename creates a sanitized filename from the URL path and current timestamp.
+// Format: v1-responses-2025-12-23T195811-a1b2c3d4.log
 //
 // Parameters:
 //   - url: The request URL
+//   - requestID: Optional request ID to include in filename
 //
 // Returns:
 //   - string: A sanitized filename for the log file
-func (l *FileRequestLogger) generateFilename(url string) string {
+func (l *FileRequestLogger) generateFilename(url string, requestID ...string) string {
 	// Extract path from URL
 	path := url
 	if strings.Contains(url, "?") {
@@ -368,12 +374,18 @@ func (l *FileRequestLogger) generateFilename(url string) string {
 	sanitized := l.sanitizeForFilename(path)
 
 	// Add timestamp
-	timestamp := time.Now().Format("2006-01-02T150405-.000000000")
-	timestamp = strings.Replace(timestamp, ".", "", -1)
+	timestamp := time.Now().Format("2006-01-02T150405")
 
-	id := requestLogID.Add(1)
+	// Use request ID if provided, otherwise use sequential ID
+	var idPart string
+	if len(requestID) > 0 && requestID[0] != "" {
+		idPart = requestID[0]
+	} else {
+		id := requestLogID.Add(1)
+		idPart = fmt.Sprintf("%d", id)
+	}
 
-	return fmt.Sprintf("%s-%s-%d.log", sanitized, timestamp, id)
+	return fmt.Sprintf("%s-%s-%s.log", sanitized, timestamp, idPart)
 }
 
 // sanitizeForFilename replaces characters that are not safe for filenames.

internal/logging/requestid.go

@@ -0,0 +1,61 @@
+package logging
+
+import (
+	"context"
+	"crypto/rand"
+	"encoding/hex"
+
+	"github.com/gin-gonic/gin"
+)
+
+// requestIDKey is the context key for storing/retrieving request IDs.
+type requestIDKey struct{}
+
+// ginRequestIDKey is the Gin context key for request IDs.
+const ginRequestIDKey = "__request_id__"
+
+// GenerateRequestID creates a new 8-character hex request ID.
+func GenerateRequestID() string {
+	b := make([]byte, 4)
+	if _, err := rand.Read(b); err != nil {
+		return "00000000"
+	}
+	return hex.EncodeToString(b)
+}
+
+// WithRequestID returns a new context with the request ID attached.
+func WithRequestID(ctx context.Context, requestID string) context.Context {
+	return context.WithValue(ctx, requestIDKey{}, requestID)
+}
+
+// GetRequestID retrieves the request ID from the context.
+// Returns empty string if not found.
+func GetRequestID(ctx context.Context) string {
+	if ctx == nil {
+		return ""
+	}
+	if id, ok := ctx.Value(requestIDKey{}).(string); ok {
+		return id
+	}
+	return ""
+}
+
+// SetGinRequestID stores the request ID in the Gin context.
+func SetGinRequestID(c *gin.Context, requestID string) {
+	if c != nil {
+		c.Set(ginRequestIDKey, requestID)
+	}
+}
+
+// GetGinRequestID retrieves the request ID from the Gin context.
+func GetGinRequestID(c *gin.Context) string {
+	if c == nil {
+		return ""
+	}
+	if id, exists := c.Get(ginRequestIDKey); exists {
+		if s, ok := id.(string); ok {
+			return s
+		}
+	}
+	return ""
+}

internal/managementasset/updater.go

@@ -24,10 +24,11 @@ import (
 )
 
 const (
-	defaultManagementReleaseURL = "https://api.github.com/repos/router-for-me/Cli-Proxy-API-Management-Center/releases/latest"
-	managementAssetName         = "management.html"
-	httpUserAgent               = "CLIProxyAPI-management-updater"
-	updateCheckInterval         = 3 * time.Hour
+	defaultManagementReleaseURL  = "https://api.github.com/repos/router-for-me/Cli-Proxy-API-Management-Center/releases/latest"
+	defaultManagementFallbackURL = "https://cpamc.router-for.me/"
+	managementAssetName          = "management.html"
+	httpUserAgent                = "CLIProxyAPI-management-updater"
+	updateCheckInterval          = 3 * time.Hour
 )
 
 // ManagementFileName exposes the control panel asset filename.
@@ -198,6 +199,16 @@ func EnsureLatestManagementHTML(ctx context.Context, staticDir string, proxyURL
 		return
 	}
 
+	localPath := filepath.Join(staticDir, managementAssetName)
+	localFileMissing := false
+	if _, errStat := os.Stat(localPath); errStat != nil {
+		if errors.Is(errStat, os.ErrNotExist) {
+			localFileMissing = true
+		} else {
+			log.WithError(errStat).Debug("failed to stat local management asset")
+		}
+	}
+
 	// Rate limiting: check only once every 3 hours
 	lastUpdateCheckMu.Lock()
 	now := time.Now()
@@ -210,15 +221,14 @@ func EnsureLatestManagementHTML(ctx context.Context, staticDir string, proxyURL
 	lastUpdateCheckTime = now
 	lastUpdateCheckMu.Unlock()
 
-	if err := os.MkdirAll(staticDir, 0o755); err != nil {
-		log.WithError(err).Warn("failed to prepare static directory for management asset")
+	if errMkdirAll := os.MkdirAll(staticDir, 0o755); errMkdirAll != nil {
+		log.WithError(errMkdirAll).Warn("failed to prepare static directory for management asset")
 		return
 	}
 
 	releaseURL := resolveReleaseURL(panelRepository)
 	client := newHTTPClient(proxyURL)
 
-	localPath := filepath.Join(staticDir, managementAssetName)
 	localHash, err := fileSHA256(localPath)
 	if err != nil {
 		if !errors.Is(err, os.ErrNotExist) {
@@ -229,6 +239,13 @@ func EnsureLatestManagementHTML(ctx context.Context, staticDir string, proxyURL
 
 	asset, remoteHash, err := fetchLatestAsset(ctx, client, releaseURL)
 	if err != nil {
+		if localFileMissing {
+			log.WithError(err).Warn("failed to fetch latest management release information, trying fallback page")
+			if ensureFallbackManagementHTML(ctx, client, localPath) {
+				return
+			}
+			return
+		}
 		log.WithError(err).Warn("failed to fetch latest management release information")
 		return
 	}
@@ -240,6 +257,13 @@ func EnsureLatestManagementHTML(ctx context.Context, staticDir string, proxyURL
 
 	data, downloadedHash, err := downloadAsset(ctx, client, asset.BrowserDownloadURL)
 	if err != nil {
+		if localFileMissing {
+			log.WithError(err).Warn("failed to download management asset, trying fallback page")
+			if ensureFallbackManagementHTML(ctx, client, localPath) {
+				return
+			}
+			return
+		}
 		log.WithError(err).Warn("failed to download management asset")
 		return
 	}
@@ -256,6 +280,22 @@ func EnsureLatestManagementHTML(ctx context.Context, staticDir string, proxyURL
 	log.Infof("management asset updated successfully (hash=%s)", downloadedHash)
 }
 
+func ensureFallbackManagementHTML(ctx context.Context, client *http.Client, localPath string) bool {
+	data, downloadedHash, err := downloadAsset(ctx, client, defaultManagementFallbackURL)
+	if err != nil {
+		log.WithError(err).Warn("failed to download fallback management control panel page")
+		return false
+	}
+
+	if err = atomicWriteFile(localPath, data); err != nil {
+		log.WithError(err).Warn("failed to persist fallback management control panel page")
+		return false
+	}
+
+	log.Infof("management asset updated from fallback page successfully (hash=%s)", downloadedHash)
+	return true
+}
+
 func resolveReleaseURL(repo string) string {
 	repo = strings.TrimSpace(repo)
 	if repo == "" {

internal/registry/model_definitions.go

@@ -727,6 +727,7 @@ func GetIFlowModels() []*ModelInfo {
 		{ID: "qwen3-max-preview", DisplayName: "Qwen3-Max-Preview", Description: "Qwen3 Max preview build", Created: 1757030400},
 		{ID: "kimi-k2-0905", DisplayName: "Kimi-K2-Instruct-0905", Description: "Moonshot Kimi K2 instruct 0905", Created: 1757030400},
 		{ID: "glm-4.6", DisplayName: "GLM-4.6", Description: "Zhipu GLM 4.6 general model", Created: 1759190400, Thinking: iFlowThinkingSupport},
+		{ID: "glm-4.7", DisplayName: "GLM-4.7", Description: "Zhipu GLM 4.7 general model", Created: 1766448000, Thinking: iFlowThinkingSupport},
 		{ID: "kimi-k2", DisplayName: "Kimi-K2", Description: "Moonshot Kimi K2 general model", Created: 1752192000},
 		{ID: "kimi-k2-thinking", DisplayName: "Kimi-K2-Thinking", Description: "Moonshot Kimi K2 thinking model", Created: 1762387200},
 		{ID: "deepseek-v3.2-chat", DisplayName: "DeepSeek-V3.2", Description: "DeepSeek V3.2 Chat", Created: 1764576000},
@@ -739,7 +740,8 @@ func GetIFlowModels() []*ModelInfo {
 		{ID: "qwen3-235b-a22b-thinking-2507", DisplayName: "Qwen3-235B-A22B-Thinking", Description: "Qwen3 235B A22B Thinking (2507)", Created: 1753401600},
 		{ID: "qwen3-235b-a22b-instruct", DisplayName: "Qwen3-235B-A22B-Instruct", Description: "Qwen3 235B A22B Instruct", Created: 1753401600},
 		{ID: "qwen3-235b", DisplayName: "Qwen3-235B-A22B", Description: "Qwen3 235B A22B", Created: 1753401600},
-		{ID: "minimax-m2", DisplayName: "MiniMax-M2", Description: "MiniMax M2", Created: 1758672000},
+		{ID: "minimax-m2", DisplayName: "MiniMax-M2", Description: "MiniMax M2", Created: 1758672000, Thinking: iFlowThinkingSupport},
+		{ID: "minimax-m2.1", DisplayName: "MiniMax-M2.1", Description: "MiniMax M2.1", Created: 1766448000, Thinking: iFlowThinkingSupport},
 	}
 	models := make([]*ModelInfo, 0, len(entries))
 	for _, entry := range entries {
@@ -771,7 +773,7 @@ func GetAntigravityModelConfig() map[string]*AntigravityModelConfig {
 	return map[string]*AntigravityModelConfig{
 		"gemini-2.5-flash":                        {Thinking: &ThinkingSupport{Min: 0, Max: 24576, ZeroAllowed: true, DynamicAllowed: true}, Name: "models/gemini-2.5-flash"},
 		"gemini-2.5-flash-lite":                   {Thinking: &ThinkingSupport{Min: 0, Max: 24576, ZeroAllowed: true, DynamicAllowed: true}, Name: "models/gemini-2.5-flash-lite"},
-		"gemini-2.5-computer-use-preview-10-2025": {Name: "models/gemini-2.5-computer-use-preview-10-2025"},
+		"gemini-2.5-computer-use-preview-10-2025": {Thinking: &ThinkingSupport{Min: 128, Max: 32768, ZeroAllowed: false, DynamicAllowed: true}, Name: "models/gemini-2.5-computer-use-preview-10-2025"},
 		"gemini-3-pro-preview":                    {Thinking: &ThinkingSupport{Min: 128, Max: 32768, ZeroAllowed: false, DynamicAllowed: true, Levels: []string{"low", "high"}}, Name: "models/gemini-3-pro-preview"},
 		"gemini-3-pro-image-preview":              {Thinking: &ThinkingSupport{Min: 128, Max: 32768, ZeroAllowed: false, DynamicAllowed: true, Levels: []string{"low", "high"}}, Name: "models/gemini-3-pro-image-preview"},
 		"gemini-3-flash-preview":                  {Thinking: &ThinkingSupport{Min: 128, Max: 32768, ZeroAllowed: false, DynamicAllowed: true, Levels: []string{"minimal", "low", "medium", "high"}}, Name: "models/gemini-3-flash-preview"},
@@ -780,6 +782,32 @@ func GetAntigravityModelConfig() map[string]*AntigravityModelConfig {
 	}
 }
 
+// LookupStaticModelInfo searches all static model definitions for a model by ID.
+// Returns nil if no matching model is found.
+func LookupStaticModelInfo(modelID string) *ModelInfo {
+	if modelID == "" {
+		return nil
+	}
+	allModels := [][]*ModelInfo{
+		GetClaudeModels(),
+		GetGeminiModels(),
+		GetGeminiVertexModels(),
+		GetGeminiCLIModels(),
+		GetAIStudioModels(),
+		GetOpenAIModels(),
+		GetQwenModels(),
+		GetIFlowModels(),
+	}
+	for _, models := range allModels {
+		for _, m := range models {
+			if m != nil && m.ID == modelID {
+				return m
+			}
+		}
+	}
+	return nil
+}
+
 // GetGitHubCopilotModels returns the available models for GitHub Copilot.
 // These models are available through the GitHub Copilot API at api.githubcopilot.com.
 func GetGitHubCopilotModels() []*ModelInfo {

internal/registry/model_registry.go

@@ -625,6 +625,131 @@ func (r *ModelRegistry) GetAvailableModels(handlerType string) []map[string]any
 	return models
 }
 
+// GetAvailableModelsByProvider returns models available for the given provider identifier.
+// Parameters:
+//   - provider: Provider identifier (e.g., "codex", "gemini", "antigravity")
+//
+// Returns:
+//   - []*ModelInfo: List of available models for the provider
+func (r *ModelRegistry) GetAvailableModelsByProvider(provider string) []*ModelInfo {
+	provider = strings.ToLower(strings.TrimSpace(provider))
+	if provider == "" {
+		return nil
+	}
+
+	r.mutex.RLock()
+	defer r.mutex.RUnlock()
+
+	type providerModel struct {
+		count int
+		info  *ModelInfo
+	}
+
+	providerModels := make(map[string]*providerModel)
+
+	for clientID, clientProvider := range r.clientProviders {
+		if clientProvider != provider {
+			continue
+		}
+		modelIDs := r.clientModels[clientID]
+		if len(modelIDs) == 0 {
+			continue
+		}
+		clientInfos := r.clientModelInfos[clientID]
+		for _, modelID := range modelIDs {
+			modelID = strings.TrimSpace(modelID)
+			if modelID == "" {
+				continue
+			}
+			entry := providerModels[modelID]
+			if entry == nil {
+				entry = &providerModel{}
+				providerModels[modelID] = entry
+			}
+			entry.count++
+			if entry.info == nil {
+				if clientInfos != nil {
+					if info := clientInfos[modelID]; info != nil {
+						entry.info = info
+					}
+				}
+				if entry.info == nil {
+					if reg, ok := r.models[modelID]; ok && reg != nil && reg.Info != nil {
+						entry.info = reg.Info
+					}
+				}
+			}
+		}
+	}
+
+	if len(providerModels) == 0 {
+		return nil
+	}
+
+	quotaExpiredDuration := 5 * time.Minute
+	now := time.Now()
+	result := make([]*ModelInfo, 0, len(providerModels))
+
+	for modelID, entry := range providerModels {
+		if entry == nil || entry.count <= 0 {
+			continue
+		}
+		registration, ok := r.models[modelID]
+
+		expiredClients := 0
+		cooldownSuspended := 0
+		otherSuspended := 0
+		if ok && registration != nil {
+			if registration.QuotaExceededClients != nil {
+				for clientID, quotaTime := range registration.QuotaExceededClients {
+					if clientID == "" {
+						continue
+					}
+					if p, okProvider := r.clientProviders[clientID]; !okProvider || p != provider {
+						continue
+					}
+					if quotaTime != nil && now.Sub(*quotaTime) < quotaExpiredDuration {
+						expiredClients++
+					}
+				}
+			}
+			if registration.SuspendedClients != nil {
+				for clientID, reason := range registration.SuspendedClients {
+					if clientID == "" {
+						continue
+					}
+					if p, okProvider := r.clientProviders[clientID]; !okProvider || p != provider {
+						continue
+					}
+					if strings.EqualFold(reason, "quota") {
+						cooldownSuspended++
+						continue
+					}
+					otherSuspended++
+				}
+			}
+		}
+
+		availableClients := entry.count
+		effectiveClients := availableClients - expiredClients - otherSuspended
+		if effectiveClients < 0 {
+			effectiveClients = 0
+		}
+
+		if effectiveClients > 0 || (availableClients > 0 && (expiredClients > 0 || cooldownSuspended > 0) && otherSuspended == 0) {
+			if entry.info != nil {
+				result = append(result, entry.info)
+				continue
+			}
+			if ok && registration != nil && registration.Info != nil {
+				result = append(result, registration.Info)
+			}
+		}
+	}
+
+	return result
+}
+
 // GetModelCount returns the number of available clients for a specific model
 // Parameters:
 //   - modelID: The model ID to check

internal/runtime/executor/aistudio_executor.go

@@ -59,6 +59,7 @@ func (e *AIStudioExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth,
 	if err != nil {
 		return resp, err
 	}
+
 	endpoint := e.buildEndpoint(req.Model, body.action, opts.Alt)
 	wsReq := &wsrelay.HTTPRequest{
 		Method:  http.MethodPost,
@@ -113,6 +114,7 @@ func (e *AIStudioExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth
 	if err != nil {
 		return nil, err
 	}
+
 	endpoint := e.buildEndpoint(req.Model, body.action, opts.Alt)
 	wsReq := &wsrelay.HTTPRequest{
 		Method:  http.MethodPost,
@@ -321,6 +323,11 @@ type translatedPayload struct {
 func (e *AIStudioExecutor) translateRequest(req cliproxyexecutor.Request, opts cliproxyexecutor.Options, stream bool) ([]byte, translatedPayload, error) {
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("gemini")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, stream)
 	payload := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), stream)
 	payload = ApplyThinkingMetadata(payload, req.Metadata, req.Model)
 	payload = util.ApplyGemini3ThinkingLevelFromMetadata(req.Model, req.Metadata, payload)
@@ -329,7 +336,7 @@ func (e *AIStudioExecutor) translateRequest(req cliproxyexecutor.Request, opts c
 	payload = util.NormalizeGeminiThinkingBudget(req.Model, payload, true)
 	payload = util.StripThinkingConfigIfUnsupported(req.Model, payload)
 	payload = fixGeminiImageAspectRatio(req.Model, payload)
-	payload = applyPayloadConfig(e.cfg, req.Model, payload)
+	payload = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", payload, originalTranslated)
 	payload, _ = sjson.DeleteBytes(payload, "generationConfig.maxOutputTokens")
 	payload, _ = sjson.DeleteBytes(payload, "generationConfig.responseMimeType")
 	payload, _ = sjson.DeleteBytes(payload, "generationConfig.responseJsonSchema")

internal/runtime/executor/antigravity_executor.go

@@ -33,18 +33,18 @@ import (
 )
 
 const (
-	antigravityBaseURLDaily = "https://daily-cloudcode-pa.sandbox.googleapis.com"
-	// antigravityBaseURLAutopush     = "https://autopush-cloudcode-pa.sandbox.googleapis.com"
-	antigravityBaseURLProd     = "https://cloudcode-pa.googleapis.com"
-	antigravityCountTokensPath = "/v1internal:countTokens"
-	antigravityStreamPath      = "/v1internal:streamGenerateContent"
-	antigravityGeneratePath    = "/v1internal:generateContent"
-	antigravityModelsPath      = "/v1internal:fetchAvailableModels"
-	antigravityClientID        = "1071006060591-tmhssin2h21lcre235vtolojh4g403ep.apps.googleusercontent.com"
-	antigravityClientSecret    = "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"
-	defaultAntigravityAgent    = "antigravity/1.11.5 windows/amd64"
-	antigravityAuthType        = "antigravity"
-	refreshSkew                = 3000 * time.Second
+	antigravityBaseURLDaily        = "https://daily-cloudcode-pa.googleapis.com"
+	antigravitySandboxBaseURLDaily = "https://daily-cloudcode-pa.sandbox.googleapis.com"
+	antigravityBaseURLProd         = "https://cloudcode-pa.googleapis.com"
+	antigravityCountTokensPath     = "/v1internal:countTokens"
+	antigravityStreamPath          = "/v1internal:streamGenerateContent"
+	antigravityGeneratePath        = "/v1internal:generateContent"
+	antigravityModelsPath          = "/v1internal:fetchAvailableModels"
+	antigravityClientID            = "1071006060591-tmhssin2h21lcre235vtolojh4g403ep.apps.googleusercontent.com"
+	antigravityClientSecret        = "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"
+	defaultAntigravityAgent        = "antigravity/1.104.0 darwin/arm64"
+	antigravityAuthType            = "antigravity"
+	refreshSkew                    = 3000 * time.Second
 )
 
 var (
@@ -76,7 +76,8 @@ func (e *AntigravityExecutor) PrepareRequest(_ *http.Request, _ *cliproxyauth.Au
 
 // Execute performs a non-streaming request to the Antigravity API.
 func (e *AntigravityExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, req cliproxyexecutor.Request, opts cliproxyexecutor.Options) (resp cliproxyexecutor.Response, err error) {
-	if strings.Contains(req.Model, "claude") {
+	isClaude := strings.Contains(strings.ToLower(req.Model), "claude")
+	if isClaude {
 		return e.executeClaudeNonStream(ctx, auth, req, opts)
 	}
 
@@ -93,13 +94,18 @@ func (e *AntigravityExecutor) Execute(ctx context.Context, auth *cliproxyauth.Au
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("antigravity")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, false)
 	translated := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
 
-	translated = applyThinkingMetadataCLI(translated, req.Metadata, req.Model)
+	translated = ApplyThinkingMetadataCLI(translated, req.Metadata, req.Model)
 	translated = util.ApplyGemini3ThinkingLevelFromMetadataCLI(req.Model, req.Metadata, translated)
-	translated = util.ApplyDefaultThinkingIfNeededCLI(req.Model, translated)
-	translated = normalizeAntigravityThinking(req.Model, translated)
-	translated = applyPayloadConfigWithRoot(e.cfg, req.Model, "antigravity", "request", translated)
+	translated = util.ApplyDefaultThinkingIfNeededCLI(req.Model, req.Metadata, translated)
+	translated = normalizeAntigravityThinking(req.Model, translated, isClaude)
+	translated = applyPayloadConfigWithRoot(e.cfg, req.Model, "antigravity", "request", translated, originalTranslated)
 
 	baseURLs := antigravityBaseURLFallbackOrder(auth)
 	httpClient := newProxyAwareHTTPClient(ctx, e.cfg, auth, 0)
@@ -188,13 +194,18 @@ func (e *AntigravityExecutor) executeClaudeNonStream(ctx context.Context, auth *
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("antigravity")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, true)
 	translated := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), true)
 
-	translated = applyThinkingMetadataCLI(translated, req.Metadata, req.Model)
+	translated = ApplyThinkingMetadataCLI(translated, req.Metadata, req.Model)
 	translated = util.ApplyGemini3ThinkingLevelFromMetadataCLI(req.Model, req.Metadata, translated)
-	translated = util.ApplyDefaultThinkingIfNeededCLI(req.Model, translated)
-	translated = normalizeAntigravityThinking(req.Model, translated)
-	translated = applyPayloadConfigWithRoot(e.cfg, req.Model, "antigravity", "request", translated)
+	translated = util.ApplyDefaultThinkingIfNeededCLI(req.Model, req.Metadata, translated)
+	translated = normalizeAntigravityThinking(req.Model, translated, true)
+	translated = applyPayloadConfigWithRoot(e.cfg, req.Model, "antigravity", "request", translated, originalTranslated)
 
 	baseURLs := antigravityBaseURLFallbackOrder(auth)
 	httpClient := newProxyAwareHTTPClient(ctx, e.cfg, auth, 0)
@@ -520,15 +531,22 @@ func (e *AntigravityExecutor) ExecuteStream(ctx context.Context, auth *cliproxya
 	reporter := newUsageReporter(ctx, e.Identifier(), req.Model, auth)
 	defer reporter.trackFailure(ctx, &err)
 
+	isClaude := strings.Contains(strings.ToLower(req.Model), "claude")
+
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("antigravity")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, true)
 	translated := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), true)
 
-	translated = applyThinkingMetadataCLI(translated, req.Metadata, req.Model)
+	translated = ApplyThinkingMetadataCLI(translated, req.Metadata, req.Model)
 	translated = util.ApplyGemini3ThinkingLevelFromMetadataCLI(req.Model, req.Metadata, translated)
-	translated = util.ApplyDefaultThinkingIfNeededCLI(req.Model, translated)
-	translated = normalizeAntigravityThinking(req.Model, translated)
-	translated = applyPayloadConfigWithRoot(e.cfg, req.Model, "antigravity", "request", translated)
+	translated = util.ApplyDefaultThinkingIfNeededCLI(req.Model, req.Metadata, translated)
+	translated = normalizeAntigravityThinking(req.Model, translated, isClaude)
+	translated = applyPayloadConfigWithRoot(e.cfg, req.Model, "antigravity", "request", translated, originalTranslated)
 
 	baseURLs := antigravityBaseURLFallbackOrder(auth)
 	httpClient := newProxyAwareHTTPClient(ctx, e.cfg, auth, 0)
@@ -676,6 +694,8 @@ func (e *AntigravityExecutor) CountTokens(ctx context.Context, auth *cliproxyaut
 	to := sdktranslator.FromString("antigravity")
 	respCtx := context.WithValue(ctx, "alt", opts.Alt)
 
+	isClaude := strings.Contains(strings.ToLower(req.Model), "claude")
+
 	baseURLs := antigravityBaseURLFallbackOrder(auth)
 	httpClient := newProxyAwareHTTPClient(ctx, e.cfg, auth, 0)
 
@@ -692,9 +712,9 @@ func (e *AntigravityExecutor) CountTokens(ctx context.Context, auth *cliproxyaut
 
 	for idx, baseURL := range baseURLs {
 		payload := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
-		payload = applyThinkingMetadataCLI(payload, req.Metadata, req.Model)
-		payload = util.ApplyDefaultThinkingIfNeededCLI(req.Model, payload)
-		payload = normalizeAntigravityThinking(req.Model, payload)
+		payload = ApplyThinkingMetadataCLI(payload, req.Metadata, req.Model)
+		payload = util.ApplyDefaultThinkingIfNeededCLI(req.Model, req.Metadata, payload)
+		payload = normalizeAntigravityThinking(req.Model, payload, isClaude)
 		payload = deleteJSONField(payload, "project")
 		payload = deleteJSONField(payload, "model")
 		payload = deleteJSONField(payload, "request.safetySettings")
@@ -1156,7 +1176,7 @@ func antigravityBaseURLFallbackOrder(auth *cliproxyauth.Auth) []string {
 	}
 	return []string{
 		antigravityBaseURLDaily,
-		// antigravityBaseURLAutopush,
+		antigravitySandboxBaseURLDaily,
 		antigravityBaseURLProd,
 	}
 }
@@ -1308,7 +1328,7 @@ func alias2ModelName(modelName string) string {
 
 // normalizeAntigravityThinking clamps or removes thinking config based on model support.
 // For Claude models, it additionally ensures thinking budget < max_tokens.
-func normalizeAntigravityThinking(model string, payload []byte) []byte {
+func normalizeAntigravityThinking(model string, payload []byte, isClaude bool) []byte {
 	payload = util.StripThinkingConfigIfUnsupported(model, payload)
 	if !util.ModelSupportsThinking(model) {
 		return payload
@@ -1320,7 +1340,6 @@ func normalizeAntigravityThinking(model string, payload []byte) []byte {
 	raw := int(budget.Int())
 	normalized := util.NormalizeThinkingBudget(model, raw)
 
-	isClaude := strings.Contains(strings.ToLower(model), "claude")
 	if isClaude {
 		effectiveMax, setDefaultMax := antigravityEffectiveMaxTokens(model, payload)
 		if effectiveMax > 0 && normalized >= effectiveMax {

internal/runtime/executor/claude_executor.go

@@ -49,33 +49,34 @@ func (e *ClaudeExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, r
 	}
 	reporter := newUsageReporter(ctx, e.Identifier(), req.Model, auth)
 	defer reporter.trackFailure(ctx, &err)
+	model := req.Model
+	if override := e.resolveUpstreamModel(req.Model, auth); override != "" {
+		model = override
+	}
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("claude")
 	// Use streaming translation to preserve function calling, except for claude.
 	stream := from != to
-	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), stream)
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
-	if upstreamModel == "" {
-		upstreamModel = req.Model
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
 	}
-	if modelOverride := e.resolveUpstreamModel(upstreamModel, auth); modelOverride != "" {
-		upstreamModel = modelOverride
-	} else if !strings.EqualFold(upstreamModel, req.Model) {
-		if modelOverride := e.resolveUpstreamModel(req.Model, auth); modelOverride != "" {
-			upstreamModel = modelOverride
-		}
-	}
-	body, _ = sjson.SetBytes(body, "model", upstreamModel)
+	originalTranslated := sdktranslator.TranslateRequest(from, to, model, originalPayload, stream)
+	body := sdktranslator.TranslateRequest(from, to, model, bytes.Clone(req.Payload), stream)
+	body, _ = sjson.SetBytes(body, "model", model)
 	// Inject thinking config based on model metadata for thinking variants
-	body = e.injectThinkingConfig(req.Model, req.Metadata, body)
+	body = e.injectThinkingConfig(model, req.Metadata, body)
 
-	if !strings.HasPrefix(upstreamModel, "claude-3-5-haiku") {
+	if !strings.HasPrefix(model, "claude-3-5-haiku") {
 		body = checkSystemInstructions(body)
 	}
-	body = applyPayloadConfig(e.cfg, req.Model, body)
+	body = applyPayloadConfigWithRoot(e.cfg, model, to.String(), "", body, originalTranslated)
+
+	// Disable thinking if tool_choice forces tool use (Anthropic API constraint)
+	body = disableThinkingIfToolChoiceForced(body)
 
 	// Ensure max_tokens > thinking.budget_tokens when thinking is enabled
-	body = ensureMaxTokensForThinking(req.Model, body)
+	body = ensureMaxTokensForThinking(model, body)
 
 	// Extract betas from body and convert to header
 	var extraBetas []string
@@ -167,26 +168,27 @@ func (e *ClaudeExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.A
 	defer reporter.trackFailure(ctx, &err)
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("claude")
-	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), true)
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
-	if upstreamModel == "" {
-		upstreamModel = req.Model
+	model := req.Model
+	if override := e.resolveUpstreamModel(req.Model, auth); override != "" {
+		model = override
 	}
-	if modelOverride := e.resolveUpstreamModel(upstreamModel, auth); modelOverride != "" {
-		upstreamModel = modelOverride
-	} else if !strings.EqualFold(upstreamModel, req.Model) {
-		if modelOverride := e.resolveUpstreamModel(req.Model, auth); modelOverride != "" {
-			upstreamModel = modelOverride
-		}
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
 	}
-	body, _ = sjson.SetBytes(body, "model", upstreamModel)
+	originalTranslated := sdktranslator.TranslateRequest(from, to, model, originalPayload, true)
+	body := sdktranslator.TranslateRequest(from, to, model, bytes.Clone(req.Payload), true)
+	body, _ = sjson.SetBytes(body, "model", model)
 	// Inject thinking config based on model metadata for thinking variants
-	body = e.injectThinkingConfig(req.Model, req.Metadata, body)
+	body = e.injectThinkingConfig(model, req.Metadata, body)
 	body = checkSystemInstructions(body)
-	body = applyPayloadConfig(e.cfg, req.Model, body)
+	body = applyPayloadConfigWithRoot(e.cfg, model, to.String(), "", body, originalTranslated)
+
+	// Disable thinking if tool_choice forces tool use (Anthropic API constraint)
+	body = disableThinkingIfToolChoiceForced(body)
 
 	// Ensure max_tokens > thinking.budget_tokens when thinking is enabled
-	body = ensureMaxTokensForThinking(req.Model, body)
+	body = ensureMaxTokensForThinking(model, body)
 
 	// Extract betas from body and convert to header
 	var extraBetas []string
@@ -310,21 +312,14 @@ func (e *ClaudeExecutor) CountTokens(ctx context.Context, auth *cliproxyauth.Aut
 	to := sdktranslator.FromString("claude")
 	// Use streaming translation to preserve function calling, except for claude.
 	stream := from != to
-	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), stream)
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
-	if upstreamModel == "" {
-		upstreamModel = req.Model
+	model := req.Model
+	if override := e.resolveUpstreamModel(req.Model, auth); override != "" {
+		model = override
 	}
-	if modelOverride := e.resolveUpstreamModel(upstreamModel, auth); modelOverride != "" {
-		upstreamModel = modelOverride
-	} else if !strings.EqualFold(upstreamModel, req.Model) {
-		if modelOverride := e.resolveUpstreamModel(req.Model, auth); modelOverride != "" {
-			upstreamModel = modelOverride
-		}
-	}
-	body, _ = sjson.SetBytes(body, "model", upstreamModel)
+	body := sdktranslator.TranslateRequest(from, to, model, bytes.Clone(req.Payload), stream)
+	body, _ = sjson.SetBytes(body, "model", model)
 
-	if !strings.HasPrefix(upstreamModel, "claude-3-5-haiku") {
+	if !strings.HasPrefix(model, "claude-3-5-haiku") {
 		body = checkSystemInstructions(body)
 	}
 
@@ -461,6 +456,19 @@ func (e *ClaudeExecutor) injectThinkingConfig(modelName string, metadata map[str
 	return util.ApplyClaudeThinkingConfig(body, budget)
 }
 
+// disableThinkingIfToolChoiceForced checks if tool_choice forces tool use and disables thinking.
+// Anthropic API does not allow thinking when tool_choice is set to "any" or a specific tool.
+// See: https://docs.anthropic.com/en/docs/build-with-claude/extended-thinking#important-considerations
+func disableThinkingIfToolChoiceForced(body []byte) []byte {
+	toolChoiceType := gjson.GetBytes(body, "tool_choice.type").String()
+	// "auto" is allowed with thinking, but "any" or "tool" (specific tool) are not
+	if toolChoiceType == "any" || toolChoiceType == "tool" {
+		// Remove thinking configuration entirely to avoid API error
+		body, _ = sjson.DeleteBytes(body, "thinking")
+	}
+	return body
+}
+
 // ensureMaxTokensForThinking ensures max_tokens > thinking.budget_tokens when thinking is enabled.
 // Anthropic API requires this constraint; violating it returns a 400 error.
 // This function should be called after all thinking configuration is finalized.
@@ -662,7 +670,14 @@ func decodeResponseBody(body io.ReadCloser, contentEncoding string) (io.ReadClos
 }
 
 func applyClaudeHeaders(r *http.Request, auth *cliproxyauth.Auth, apiKey string, stream bool, extraBetas []string) {
-	r.Header.Set("Authorization", "Bearer "+apiKey)
+	useAPIKey := auth != nil && auth.Attributes != nil && strings.TrimSpace(auth.Attributes["api_key"]) != ""
+	isAnthropicBase := r.URL != nil && strings.EqualFold(r.URL.Scheme, "https") && strings.EqualFold(r.URL.Host, "api.anthropic.com")
+	if isAnthropicBase && useAPIKey {
+		r.Header.Del("Authorization")
+		r.Header.Set("x-api-key", apiKey)
+	} else {
+		r.Header.Set("Authorization", "Bearer "+apiKey)
+	}
 	r.Header.Set("Content-Type", "application/json")
 
 	var ginHeaders http.Header

internal/runtime/executor/codex_executor.go

@@ -49,18 +49,26 @@ func (e *CodexExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, re
 	reporter := newUsageReporter(ctx, e.Identifier(), req.Model, auth)
 	defer reporter.trackFailure(ctx, &err)
 
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
+	model := req.Model
+	if override := e.resolveUpstreamModel(req.Model, auth); override != "" {
+		model = override
+	}
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("codex")
-	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
-	body = ApplyReasoningEffortMetadata(body, req.Metadata, req.Model, "reasoning.effort", false)
-	body = NormalizeThinkingConfig(body, upstreamModel, false)
-	if errValidate := ValidateThinkingConfig(body, upstreamModel); errValidate != nil {
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, model, originalPayload, false)
+	body := sdktranslator.TranslateRequest(from, to, model, bytes.Clone(req.Payload), false)
+	body = ApplyReasoningEffortMetadata(body, req.Metadata, model, "reasoning.effort", false)
+	body = NormalizeThinkingConfig(body, model, false)
+	if errValidate := ValidateThinkingConfig(body, model); errValidate != nil {
 		return resp, errValidate
 	}
-	body = applyPayloadConfig(e.cfg, req.Model, body)
-	body, _ = sjson.SetBytes(body, "model", upstreamModel)
+	body = applyPayloadConfigWithRoot(e.cfg, model, to.String(), "", body, originalTranslated)
+	body, _ = sjson.SetBytes(body, "model", model)
 	body, _ = sjson.SetBytes(body, "stream", true)
 	body, _ = sjson.DeleteBytes(body, "previous_response_id")
 
@@ -146,20 +154,28 @@ func (e *CodexExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.Au
 	reporter := newUsageReporter(ctx, e.Identifier(), req.Model, auth)
 	defer reporter.trackFailure(ctx, &err)
 
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
+	model := req.Model
+	if override := e.resolveUpstreamModel(req.Model, auth); override != "" {
+		model = override
+	}
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("codex")
-	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), true)
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, model, originalPayload, true)
+	body := sdktranslator.TranslateRequest(from, to, model, bytes.Clone(req.Payload), true)
 
-	body = ApplyReasoningEffortMetadata(body, req.Metadata, req.Model, "reasoning.effort", false)
-	body = NormalizeThinkingConfig(body, upstreamModel, false)
-	if errValidate := ValidateThinkingConfig(body, upstreamModel); errValidate != nil {
+	body = ApplyReasoningEffortMetadata(body, req.Metadata, model, "reasoning.effort", false)
+	body = NormalizeThinkingConfig(body, model, false)
+	if errValidate := ValidateThinkingConfig(body, model); errValidate != nil {
 		return nil, errValidate
 	}
-	body = applyPayloadConfig(e.cfg, req.Model, body)
+	body = applyPayloadConfigWithRoot(e.cfg, model, to.String(), "", body, originalTranslated)
 	body, _ = sjson.DeleteBytes(body, "previous_response_id")
-	body, _ = sjson.SetBytes(body, "model", upstreamModel)
+	body, _ = sjson.SetBytes(body, "model", model)
 
 	url := strings.TrimSuffix(baseURL, "/") + "/responses"
 	httpReq, err := e.cacheHelper(ctx, from, url, req, body)
@@ -246,20 +262,21 @@ func (e *CodexExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.Au
 }
 
 func (e *CodexExecutor) CountTokens(ctx context.Context, auth *cliproxyauth.Auth, req cliproxyexecutor.Request, opts cliproxyexecutor.Options) (cliproxyexecutor.Response, error) {
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
+	model := req.Model
+	if override := e.resolveUpstreamModel(req.Model, auth); override != "" {
+		model = override
+	}
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("codex")
-	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
+	body := sdktranslator.TranslateRequest(from, to, model, bytes.Clone(req.Payload), false)
 
-	modelForCounting := req.Model
-
-	body = ApplyReasoningEffortMetadata(body, req.Metadata, req.Model, "reasoning.effort", false)
-	body, _ = sjson.SetBytes(body, "model", upstreamModel)
+	body = ApplyReasoningEffortMetadata(body, req.Metadata, model, "reasoning.effort", false)
+	body, _ = sjson.SetBytes(body, "model", model)
 	body, _ = sjson.DeleteBytes(body, "previous_response_id")
 	body, _ = sjson.SetBytes(body, "stream", false)
 
-	enc, err := tokenizerForCodexModel(modelForCounting)
+	enc, err := tokenizerForCodexModel(model)
 	if err != nil {
 		return cliproxyexecutor.Response{}, fmt.Errorf("codex executor: tokenizer init failed: %w", err)
 	}
@@ -520,3 +537,87 @@ func codexCreds(a *cliproxyauth.Auth) (apiKey, baseURL string) {
 	}
 	return
 }
+
+func (e *CodexExecutor) resolveUpstreamModel(alias string, auth *cliproxyauth.Auth) string {
+	trimmed := strings.TrimSpace(alias)
+	if trimmed == "" {
+		return ""
+	}
+
+	entry := e.resolveCodexConfig(auth)
+	if entry == nil {
+		return ""
+	}
+
+	normalizedModel, metadata := util.NormalizeThinkingModel(trimmed)
+
+	// Candidate names to match against configured aliases/names.
+	candidates := []string{strings.TrimSpace(normalizedModel)}
+	if !strings.EqualFold(normalizedModel, trimmed) {
+		candidates = append(candidates, trimmed)
+	}
+	if original := util.ResolveOriginalModel(normalizedModel, metadata); original != "" && !strings.EqualFold(original, normalizedModel) {
+		candidates = append(candidates, original)
+	}
+
+	for i := range entry.Models {
+		model := entry.Models[i]
+		name := strings.TrimSpace(model.Name)
+		modelAlias := strings.TrimSpace(model.Alias)
+
+		for _, candidate := range candidates {
+			if candidate == "" {
+				continue
+			}
+			if modelAlias != "" && strings.EqualFold(modelAlias, candidate) {
+				if name != "" {
+					return name
+				}
+				return candidate
+			}
+			if name != "" && strings.EqualFold(name, candidate) {
+				return name
+			}
+		}
+	}
+	return ""
+}
+
+func (e *CodexExecutor) resolveCodexConfig(auth *cliproxyauth.Auth) *config.CodexKey {
+	if auth == nil || e.cfg == nil {
+		return nil
+	}
+	var attrKey, attrBase string
+	if auth.Attributes != nil {
+		attrKey = strings.TrimSpace(auth.Attributes["api_key"])
+		attrBase = strings.TrimSpace(auth.Attributes["base_url"])
+	}
+	for i := range e.cfg.CodexKey {
+		entry := &e.cfg.CodexKey[i]
+		cfgKey := strings.TrimSpace(entry.APIKey)
+		cfgBase := strings.TrimSpace(entry.BaseURL)
+		if attrKey != "" && attrBase != "" {
+			if strings.EqualFold(cfgKey, attrKey) && strings.EqualFold(cfgBase, attrBase) {
+				return entry
+			}
+			continue
+		}
+		if attrKey != "" && strings.EqualFold(cfgKey, attrKey) {
+			if cfgBase == "" || strings.EqualFold(cfgBase, attrBase) {
+				return entry
+			}
+		}
+		if attrKey == "" && attrBase != "" && strings.EqualFold(cfgBase, attrBase) {
+			return entry
+		}
+	}
+	if attrKey != "" {
+		for i := range e.cfg.CodexKey {
+			entry := &e.cfg.CodexKey[i]
+			if strings.EqualFold(strings.TrimSpace(entry.APIKey), attrKey) {
+				return entry
+			}
+		}
+	}
+	return nil
+}

internal/runtime/executor/gemini_cli_executor.go

@@ -77,14 +77,19 @@ func (e *GeminiCLIExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("gemini-cli")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, false)
 	basePayload := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
-	basePayload = applyThinkingMetadataCLI(basePayload, req.Metadata, req.Model)
+	basePayload = ApplyThinkingMetadataCLI(basePayload, req.Metadata, req.Model)
 	basePayload = util.ApplyGemini3ThinkingLevelFromMetadataCLI(req.Model, req.Metadata, basePayload)
-	basePayload = util.ApplyDefaultThinkingIfNeededCLI(req.Model, basePayload)
+	basePayload = util.ApplyDefaultThinkingIfNeededCLI(req.Model, req.Metadata, basePayload)
 	basePayload = util.NormalizeGeminiCLIThinkingBudget(req.Model, basePayload)
 	basePayload = util.StripThinkingConfigIfUnsupported(req.Model, basePayload)
 	basePayload = fixGeminiCLIImageAspectRatio(req.Model, basePayload)
-	basePayload = applyPayloadConfigWithRoot(e.cfg, req.Model, "gemini", "request", basePayload)
+	basePayload = applyPayloadConfigWithRoot(e.cfg, req.Model, "gemini", "request", basePayload, originalTranslated)
 
 	action := "generateContent"
 	if req.Metadata != nil {
@@ -216,14 +221,19 @@ func (e *GeminiCLIExecutor) ExecuteStream(ctx context.Context, auth *cliproxyaut
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("gemini-cli")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, true)
 	basePayload := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), true)
-	basePayload = applyThinkingMetadataCLI(basePayload, req.Metadata, req.Model)
+	basePayload = ApplyThinkingMetadataCLI(basePayload, req.Metadata, req.Model)
 	basePayload = util.ApplyGemini3ThinkingLevelFromMetadataCLI(req.Model, req.Metadata, basePayload)
-	basePayload = util.ApplyDefaultThinkingIfNeededCLI(req.Model, basePayload)
+	basePayload = util.ApplyDefaultThinkingIfNeededCLI(req.Model, req.Metadata, basePayload)
 	basePayload = util.NormalizeGeminiCLIThinkingBudget(req.Model, basePayload)
 	basePayload = util.StripThinkingConfigIfUnsupported(req.Model, basePayload)
 	basePayload = fixGeminiCLIImageAspectRatio(req.Model, basePayload)
-	basePayload = applyPayloadConfigWithRoot(e.cfg, req.Model, "gemini", "request", basePayload)
+	basePayload = applyPayloadConfigWithRoot(e.cfg, req.Model, "gemini", "request", basePayload, originalTranslated)
 
 	projectID := resolveGeminiProjectID(auth)
 
@@ -318,7 +328,7 @@ func (e *GeminiCLIExecutor) ExecuteStream(ctx context.Context, auth *cliproxyaut
 
 		out := make(chan cliproxyexecutor.StreamChunk)
 		stream = out
-		go func(resp *http.Response, reqBody []byte, attempt string) {
+		go func(resp *http.Response, reqBody []byte, attemptModel string) {
 			defer close(out)
 			defer func() {
 				if errClose := resp.Body.Close(); errClose != nil {
@@ -336,14 +346,14 @@ func (e *GeminiCLIExecutor) ExecuteStream(ctx context.Context, auth *cliproxyaut
 						reporter.publish(ctx, detail)
 					}
 					if bytes.HasPrefix(line, dataTag) {
-						segments := sdktranslator.TranslateStream(respCtx, to, from, attempt, bytes.Clone(opts.OriginalRequest), reqBody, bytes.Clone(line), &param)
+						segments := sdktranslator.TranslateStream(respCtx, to, from, attemptModel, bytes.Clone(opts.OriginalRequest), reqBody, bytes.Clone(line), &param)
 						for i := range segments {
 							out <- cliproxyexecutor.StreamChunk{Payload: []byte(segments[i])}
 						}
 					}
 				}
 
-				segments := sdktranslator.TranslateStream(respCtx, to, from, attempt, bytes.Clone(opts.OriginalRequest), reqBody, bytes.Clone([]byte("[DONE]")), &param)
+				segments := sdktranslator.TranslateStream(respCtx, to, from, attemptModel, bytes.Clone(opts.OriginalRequest), reqBody, bytes.Clone([]byte("[DONE]")), &param)
 				for i := range segments {
 					out <- cliproxyexecutor.StreamChunk{Payload: []byte(segments[i])}
 				}
@@ -365,12 +375,12 @@ func (e *GeminiCLIExecutor) ExecuteStream(ctx context.Context, auth *cliproxyaut
 			appendAPIResponseChunk(ctx, e.cfg, data)
 			reporter.publish(ctx, parseGeminiCLIUsage(data))
 			var param any
-			segments := sdktranslator.TranslateStream(respCtx, to, from, attempt, bytes.Clone(opts.OriginalRequest), reqBody, data, &param)
+			segments := sdktranslator.TranslateStream(respCtx, to, from, attemptModel, bytes.Clone(opts.OriginalRequest), reqBody, data, &param)
 			for i := range segments {
 				out <- cliproxyexecutor.StreamChunk{Payload: []byte(segments[i])}
 			}
 
-			segments = sdktranslator.TranslateStream(respCtx, to, from, attempt, bytes.Clone(opts.OriginalRequest), reqBody, bytes.Clone([]byte("[DONE]")), &param)
+			segments = sdktranslator.TranslateStream(respCtx, to, from, attemptModel, bytes.Clone(opts.OriginalRequest), reqBody, bytes.Clone([]byte("[DONE]")), &param)
 			for i := range segments {
 				out <- cliproxyexecutor.StreamChunk{Payload: []byte(segments[i])}
 			}
@@ -417,15 +427,17 @@ func (e *GeminiCLIExecutor) CountTokens(ctx context.Context, auth *cliproxyauth.
 	var lastStatus int
 	var lastBody []byte
 
+	// The loop variable attemptModel is only used as the concrete model id sent to the upstream
+	// Gemini CLI endpoint when iterating fallback variants.
 	for _, attemptModel := range models {
 		payload := sdktranslator.TranslateRequest(from, to, attemptModel, bytes.Clone(req.Payload), false)
-		payload = applyThinkingMetadataCLI(payload, req.Metadata, req.Model)
+		payload = ApplyThinkingMetadataCLI(payload, req.Metadata, req.Model)
 		payload = util.ApplyGemini3ThinkingLevelFromMetadataCLI(req.Model, req.Metadata, payload)
 		payload = deleteJSONField(payload, "project")
 		payload = deleteJSONField(payload, "model")
 		payload = deleteJSONField(payload, "request.safetySettings")
 		payload = util.StripThinkingConfigIfUnsupported(req.Model, payload)
-		payload = fixGeminiCLIImageAspectRatio(attemptModel, payload)
+		payload = fixGeminiCLIImageAspectRatio(req.Model, payload)
 
 		tok, errTok := tokenSource.Token()
 		if errTok != nil {

internal/runtime/executor/gemini_executor.go

@@ -77,19 +77,27 @@ func (e *GeminiExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, r
 	reporter := newUsageReporter(ctx, e.Identifier(), req.Model, auth)
 	defer reporter.trackFailure(ctx, &err)
 
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
+	model := req.Model
+	if override := e.resolveUpstreamModel(model, auth); override != "" {
+		model = override
+	}
 
 	// Official Gemini API via API key or OAuth bearer
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("gemini")
-	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
-	body = ApplyThinkingMetadata(body, req.Metadata, req.Model)
-	body = util.ApplyDefaultThinkingIfNeeded(req.Model, body)
-	body = util.NormalizeGeminiThinkingBudget(req.Model, body)
-	body = util.StripThinkingConfigIfUnsupported(req.Model, body)
-	body = fixGeminiImageAspectRatio(req.Model, body)
-	body = applyPayloadConfig(e.cfg, req.Model, body)
-	body, _ = sjson.SetBytes(body, "model", upstreamModel)
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, model, originalPayload, false)
+	body := sdktranslator.TranslateRequest(from, to, model, bytes.Clone(req.Payload), false)
+	body = ApplyThinkingMetadata(body, req.Metadata, model)
+	body = util.ApplyDefaultThinkingIfNeeded(model, body)
+	body = util.NormalizeGeminiThinkingBudget(model, body)
+	body = util.StripThinkingConfigIfUnsupported(model, body)
+	body = fixGeminiImageAspectRatio(model, body)
+	body = applyPayloadConfigWithRoot(e.cfg, model, to.String(), "", body, originalTranslated)
+	body, _ = sjson.SetBytes(body, "model", model)
 
 	action := "generateContent"
 	if req.Metadata != nil {
@@ -98,7 +106,7 @@ func (e *GeminiExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, r
 		}
 	}
 	baseURL := resolveGeminiBaseURL(auth)
-	url := fmt.Sprintf("%s/%s/models/%s:%s", baseURL, glAPIVersion, upstreamModel, action)
+	url := fmt.Sprintf("%s/%s/models/%s:%s", baseURL, glAPIVersion, model, action)
 	if opts.Alt != "" && action != "countTokens" {
 		url = url + fmt.Sprintf("?$alt=%s", opts.Alt)
 	}
@@ -173,21 +181,29 @@ func (e *GeminiExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.A
 	reporter := newUsageReporter(ctx, e.Identifier(), req.Model, auth)
 	defer reporter.trackFailure(ctx, &err)
 
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
+	model := req.Model
+	if override := e.resolveUpstreamModel(model, auth); override != "" {
+		model = override
+	}
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("gemini")
-	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), true)
-	body = ApplyThinkingMetadata(body, req.Metadata, req.Model)
-	body = util.ApplyDefaultThinkingIfNeeded(req.Model, body)
-	body = util.NormalizeGeminiThinkingBudget(req.Model, body)
-	body = util.StripThinkingConfigIfUnsupported(req.Model, body)
-	body = fixGeminiImageAspectRatio(req.Model, body)
-	body = applyPayloadConfig(e.cfg, req.Model, body)
-	body, _ = sjson.SetBytes(body, "model", upstreamModel)
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, model, originalPayload, true)
+	body := sdktranslator.TranslateRequest(from, to, model, bytes.Clone(req.Payload), true)
+	body = ApplyThinkingMetadata(body, req.Metadata, model)
+	body = util.ApplyDefaultThinkingIfNeeded(model, body)
+	body = util.NormalizeGeminiThinkingBudget(model, body)
+	body = util.StripThinkingConfigIfUnsupported(model, body)
+	body = fixGeminiImageAspectRatio(model, body)
+	body = applyPayloadConfigWithRoot(e.cfg, model, to.String(), "", body, originalTranslated)
+	body, _ = sjson.SetBytes(body, "model", model)
 
 	baseURL := resolveGeminiBaseURL(auth)
-	url := fmt.Sprintf("%s/%s/models/%s:%s", baseURL, glAPIVersion, upstreamModel, "streamGenerateContent")
+	url := fmt.Sprintf("%s/%s/models/%s:%s", baseURL, glAPIVersion, model, "streamGenerateContent")
 	if opts.Alt == "" {
 		url = url + "?alt=sse"
 	} else {
@@ -287,19 +303,25 @@ func (e *GeminiExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.A
 func (e *GeminiExecutor) CountTokens(ctx context.Context, auth *cliproxyauth.Auth, req cliproxyexecutor.Request, opts cliproxyexecutor.Options) (cliproxyexecutor.Response, error) {
 	apiKey, bearer := geminiCreds(auth)
 
+	model := req.Model
+	if override := e.resolveUpstreamModel(model, auth); override != "" {
+		model = override
+	}
+
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("gemini")
-	translatedReq := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
-	translatedReq = ApplyThinkingMetadata(translatedReq, req.Metadata, req.Model)
-	translatedReq = util.StripThinkingConfigIfUnsupported(req.Model, translatedReq)
-	translatedReq = fixGeminiImageAspectRatio(req.Model, translatedReq)
+	translatedReq := sdktranslator.TranslateRequest(from, to, model, bytes.Clone(req.Payload), false)
+	translatedReq = ApplyThinkingMetadata(translatedReq, req.Metadata, model)
+	translatedReq = util.StripThinkingConfigIfUnsupported(model, translatedReq)
+	translatedReq = fixGeminiImageAspectRatio(model, translatedReq)
 	respCtx := context.WithValue(ctx, "alt", opts.Alt)
 	translatedReq, _ = sjson.DeleteBytes(translatedReq, "tools")
 	translatedReq, _ = sjson.DeleteBytes(translatedReq, "generationConfig")
 	translatedReq, _ = sjson.DeleteBytes(translatedReq, "safetySettings")
+	translatedReq, _ = sjson.SetBytes(translatedReq, "model", model)
 
 	baseURL := resolveGeminiBaseURL(auth)
-	url := fmt.Sprintf("%s/%s/models/%s:%s", baseURL, glAPIVersion, req.Model, "countTokens")
+	url := fmt.Sprintf("%s/%s/models/%s:%s", baseURL, glAPIVersion, model, "countTokens")
 
 	requestBody := bytes.NewReader(translatedReq)
 
@@ -398,6 +420,90 @@ func resolveGeminiBaseURL(auth *cliproxyauth.Auth) string {
 	return base
 }
 
+func (e *GeminiExecutor) resolveUpstreamModel(alias string, auth *cliproxyauth.Auth) string {
+	trimmed := strings.TrimSpace(alias)
+	if trimmed == "" {
+		return ""
+	}
+
+	entry := e.resolveGeminiConfig(auth)
+	if entry == nil {
+		return ""
+	}
+
+	normalizedModel, metadata := util.NormalizeThinkingModel(trimmed)
+
+	// Candidate names to match against configured aliases/names.
+	candidates := []string{strings.TrimSpace(normalizedModel)}
+	if !strings.EqualFold(normalizedModel, trimmed) {
+		candidates = append(candidates, trimmed)
+	}
+	if original := util.ResolveOriginalModel(normalizedModel, metadata); original != "" && !strings.EqualFold(original, normalizedModel) {
+		candidates = append(candidates, original)
+	}
+
+	for i := range entry.Models {
+		model := entry.Models[i]
+		name := strings.TrimSpace(model.Name)
+		modelAlias := strings.TrimSpace(model.Alias)
+
+		for _, candidate := range candidates {
+			if candidate == "" {
+				continue
+			}
+			if modelAlias != "" && strings.EqualFold(modelAlias, candidate) {
+				if name != "" {
+					return name
+				}
+				return candidate
+			}
+			if name != "" && strings.EqualFold(name, candidate) {
+				return name
+			}
+		}
+	}
+	return ""
+}
+
+func (e *GeminiExecutor) resolveGeminiConfig(auth *cliproxyauth.Auth) *config.GeminiKey {
+	if auth == nil || e.cfg == nil {
+		return nil
+	}
+	var attrKey, attrBase string
+	if auth.Attributes != nil {
+		attrKey = strings.TrimSpace(auth.Attributes["api_key"])
+		attrBase = strings.TrimSpace(auth.Attributes["base_url"])
+	}
+	for i := range e.cfg.GeminiKey {
+		entry := &e.cfg.GeminiKey[i]
+		cfgKey := strings.TrimSpace(entry.APIKey)
+		cfgBase := strings.TrimSpace(entry.BaseURL)
+		if attrKey != "" && attrBase != "" {
+			if strings.EqualFold(cfgKey, attrKey) && strings.EqualFold(cfgBase, attrBase) {
+				return entry
+			}
+			continue
+		}
+		if attrKey != "" && strings.EqualFold(cfgKey, attrKey) {
+			if cfgBase == "" || strings.EqualFold(cfgBase, attrBase) {
+				return entry
+			}
+		}
+		if attrKey == "" && attrBase != "" && strings.EqualFold(cfgBase, attrBase) {
+			return entry
+		}
+	}
+	if attrKey != "" {
+		for i := range e.cfg.GeminiKey {
+			entry := &e.cfg.GeminiKey[i]
+			if strings.EqualFold(strings.TrimSpace(entry.APIKey), attrKey) {
+				return entry
+			}
+		}
+	}
+	return nil
+}
+
 func applyGeminiHeaders(req *http.Request, auth *cliproxyauth.Auth) {
 	var attrs map[string]string
 	if auth != nil {

internal/runtime/executor/gemini_vertex_executor.go

@@ -120,10 +120,13 @@ func (e *GeminiVertexExecutor) executeWithServiceAccount(ctx context.Context, au
 	reporter := newUsageReporter(ctx, e.Identifier(), req.Model, auth)
 	defer reporter.trackFailure(ctx, &err)
 
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
-
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("gemini")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, false)
 	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
 	if budgetOverride, includeOverride, ok := util.ResolveThinkingConfigFromMetadata(req.Model, req.Metadata); ok && util.ModelSupportsThinking(req.Model) {
 		if budgetOverride != nil {
@@ -136,8 +139,8 @@ func (e *GeminiVertexExecutor) executeWithServiceAccount(ctx context.Context, au
 	body = util.NormalizeGeminiThinkingBudget(req.Model, body)
 	body = util.StripThinkingConfigIfUnsupported(req.Model, body)
 	body = fixGeminiImageAspectRatio(req.Model, body)
-	body = applyPayloadConfig(e.cfg, req.Model, body)
-	body, _ = sjson.SetBytes(body, "model", upstreamModel)
+	body = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", body, originalTranslated)
+	body, _ = sjson.SetBytes(body, "model", req.Model)
 
 	action := "generateContent"
 	if req.Metadata != nil {
@@ -146,7 +149,7 @@ func (e *GeminiVertexExecutor) executeWithServiceAccount(ctx context.Context, au
 		}
 	}
 	baseURL := vertexBaseURL(location)
-	url := fmt.Sprintf("%s/%s/projects/%s/locations/%s/publishers/google/models/%s:%s", baseURL, vertexAPIVersion, projectID, location, upstreamModel, action)
+	url := fmt.Sprintf("%s/%s/projects/%s/locations/%s/publishers/google/models/%s:%s", baseURL, vertexAPIVersion, projectID, location, req.Model, action)
 	if opts.Alt != "" && action != "countTokens" {
 		url = url + fmt.Sprintf("?$alt=%s", opts.Alt)
 	}
@@ -220,24 +223,32 @@ func (e *GeminiVertexExecutor) executeWithAPIKey(ctx context.Context, auth *clip
 	reporter := newUsageReporter(ctx, e.Identifier(), req.Model, auth)
 	defer reporter.trackFailure(ctx, &err)
 
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
+	model := req.Model
+	if override := e.resolveUpstreamModel(req.Model, auth); override != "" {
+		model = override
+	}
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("gemini")
-	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
-	if budgetOverride, includeOverride, ok := util.ResolveThinkingConfigFromMetadata(req.Model, req.Metadata); ok && util.ModelSupportsThinking(req.Model) {
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, model, originalPayload, false)
+	body := sdktranslator.TranslateRequest(from, to, model, bytes.Clone(req.Payload), false)
+	if budgetOverride, includeOverride, ok := util.ResolveThinkingConfigFromMetadata(model, req.Metadata); ok && util.ModelSupportsThinking(model) {
 		if budgetOverride != nil {
-			norm := util.NormalizeThinkingBudget(req.Model, *budgetOverride)
+			norm := util.NormalizeThinkingBudget(model, *budgetOverride)
 			budgetOverride = &norm
 		}
 		body = util.ApplyGeminiThinkingConfig(body, budgetOverride, includeOverride)
 	}
-	body = util.ApplyDefaultThinkingIfNeeded(req.Model, body)
-	body = util.NormalizeGeminiThinkingBudget(req.Model, body)
-	body = util.StripThinkingConfigIfUnsupported(req.Model, body)
-	body = fixGeminiImageAspectRatio(req.Model, body)
-	body = applyPayloadConfig(e.cfg, req.Model, body)
-	body, _ = sjson.SetBytes(body, "model", upstreamModel)
+	body = util.ApplyDefaultThinkingIfNeeded(model, body)
+	body = util.NormalizeGeminiThinkingBudget(model, body)
+	body = util.StripThinkingConfigIfUnsupported(model, body)
+	body = fixGeminiImageAspectRatio(model, body)
+	body = applyPayloadConfigWithRoot(e.cfg, model, to.String(), "", body, originalTranslated)
+	body, _ = sjson.SetBytes(body, "model", model)
 
 	action := "generateContent"
 	if req.Metadata != nil {
@@ -250,7 +261,7 @@ func (e *GeminiVertexExecutor) executeWithAPIKey(ctx context.Context, auth *clip
 	if baseURL == "" {
 		baseURL = "https://generativelanguage.googleapis.com"
 	}
-	url := fmt.Sprintf("%s/%s/publishers/google/models/%s:%s", baseURL, vertexAPIVersion, upstreamModel, action)
+	url := fmt.Sprintf("%s/%s/publishers/google/models/%s:%s", baseURL, vertexAPIVersion, model, action)
 	if opts.Alt != "" && action != "countTokens" {
 		url = url + fmt.Sprintf("?$alt=%s", opts.Alt)
 	}
@@ -321,10 +332,13 @@ func (e *GeminiVertexExecutor) executeStreamWithServiceAccount(ctx context.Conte
 	reporter := newUsageReporter(ctx, e.Identifier(), req.Model, auth)
 	defer reporter.trackFailure(ctx, &err)
 
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
-
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("gemini")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, true)
 	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), true)
 	if budgetOverride, includeOverride, ok := util.ResolveThinkingConfigFromMetadata(req.Model, req.Metadata); ok && util.ModelSupportsThinking(req.Model) {
 		if budgetOverride != nil {
@@ -337,11 +351,11 @@ func (e *GeminiVertexExecutor) executeStreamWithServiceAccount(ctx context.Conte
 	body = util.NormalizeGeminiThinkingBudget(req.Model, body)
 	body = util.StripThinkingConfigIfUnsupported(req.Model, body)
 	body = fixGeminiImageAspectRatio(req.Model, body)
-	body = applyPayloadConfig(e.cfg, req.Model, body)
-	body, _ = sjson.SetBytes(body, "model", upstreamModel)
+	body = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", body, originalTranslated)
+	body, _ = sjson.SetBytes(body, "model", req.Model)
 
 	baseURL := vertexBaseURL(location)
-	url := fmt.Sprintf("%s/%s/projects/%s/locations/%s/publishers/google/models/%s:%s", baseURL, vertexAPIVersion, projectID, location, upstreamModel, "streamGenerateContent")
+	url := fmt.Sprintf("%s/%s/projects/%s/locations/%s/publishers/google/models/%s:%s", baseURL, vertexAPIVersion, projectID, location, req.Model, "streamGenerateContent")
 	if opts.Alt == "" {
 		url = url + "?alt=sse"
 	} else {
@@ -438,30 +452,38 @@ func (e *GeminiVertexExecutor) executeStreamWithAPIKey(ctx context.Context, auth
 	reporter := newUsageReporter(ctx, e.Identifier(), req.Model, auth)
 	defer reporter.trackFailure(ctx, &err)
 
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
+	model := req.Model
+	if override := e.resolveUpstreamModel(req.Model, auth); override != "" {
+		model = override
+	}
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("gemini")
-	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), true)
-	if budgetOverride, includeOverride, ok := util.ResolveThinkingConfigFromMetadata(req.Model, req.Metadata); ok && util.ModelSupportsThinking(req.Model) {
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, model, originalPayload, true)
+	body := sdktranslator.TranslateRequest(from, to, model, bytes.Clone(req.Payload), true)
+	if budgetOverride, includeOverride, ok := util.ResolveThinkingConfigFromMetadata(model, req.Metadata); ok && util.ModelSupportsThinking(model) {
 		if budgetOverride != nil {
-			norm := util.NormalizeThinkingBudget(req.Model, *budgetOverride)
+			norm := util.NormalizeThinkingBudget(model, *budgetOverride)
 			budgetOverride = &norm
 		}
 		body = util.ApplyGeminiThinkingConfig(body, budgetOverride, includeOverride)
 	}
-	body = util.ApplyDefaultThinkingIfNeeded(req.Model, body)
-	body = util.NormalizeGeminiThinkingBudget(req.Model, body)
-	body = util.StripThinkingConfigIfUnsupported(req.Model, body)
-	body = fixGeminiImageAspectRatio(req.Model, body)
-	body = applyPayloadConfig(e.cfg, req.Model, body)
-	body, _ = sjson.SetBytes(body, "model", upstreamModel)
+	body = util.ApplyDefaultThinkingIfNeeded(model, body)
+	body = util.NormalizeGeminiThinkingBudget(model, body)
+	body = util.StripThinkingConfigIfUnsupported(model, body)
+	body = fixGeminiImageAspectRatio(model, body)
+	body = applyPayloadConfigWithRoot(e.cfg, model, to.String(), "", body, originalTranslated)
+	body, _ = sjson.SetBytes(body, "model", model)
 
 	// For API key auth, use simpler URL format without project/location
 	if baseURL == "" {
 		baseURL = "https://generativelanguage.googleapis.com"
 	}
-	url := fmt.Sprintf("%s/%s/publishers/google/models/%s:%s", baseURL, vertexAPIVersion, upstreamModel, "streamGenerateContent")
+	url := fmt.Sprintf("%s/%s/publishers/google/models/%s:%s", baseURL, vertexAPIVersion, model, "streamGenerateContent")
 	if opts.Alt == "" {
 		url = url + "?alt=sse"
 	} else {
@@ -552,8 +574,6 @@ func (e *GeminiVertexExecutor) executeStreamWithAPIKey(ctx context.Context, auth
 
 // countTokensWithServiceAccount counts tokens using service account credentials.
 func (e *GeminiVertexExecutor) countTokensWithServiceAccount(ctx context.Context, auth *cliproxyauth.Auth, req cliproxyexecutor.Request, opts cliproxyexecutor.Options, projectID, location string, saJSON []byte) (cliproxyexecutor.Response, error) {
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
-
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("gemini")
 	translatedReq := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
@@ -566,14 +586,14 @@ func (e *GeminiVertexExecutor) countTokensWithServiceAccount(ctx context.Context
 	}
 	translatedReq = util.StripThinkingConfigIfUnsupported(req.Model, translatedReq)
 	translatedReq = fixGeminiImageAspectRatio(req.Model, translatedReq)
-	translatedReq, _ = sjson.SetBytes(translatedReq, "model", upstreamModel)
+	translatedReq, _ = sjson.SetBytes(translatedReq, "model", req.Model)
 	respCtx := context.WithValue(ctx, "alt", opts.Alt)
 	translatedReq, _ = sjson.DeleteBytes(translatedReq, "tools")
 	translatedReq, _ = sjson.DeleteBytes(translatedReq, "generationConfig")
 	translatedReq, _ = sjson.DeleteBytes(translatedReq, "safetySettings")
 
 	baseURL := vertexBaseURL(location)
-	url := fmt.Sprintf("%s/%s/projects/%s/locations/%s/publishers/google/models/%s:%s", baseURL, vertexAPIVersion, projectID, location, upstreamModel, "countTokens")
+	url := fmt.Sprintf("%s/%s/projects/%s/locations/%s/publishers/google/models/%s:%s", baseURL, vertexAPIVersion, projectID, location, req.Model, "countTokens")
 
 	httpReq, errNewReq := http.NewRequestWithContext(respCtx, http.MethodPost, url, bytes.NewReader(translatedReq))
 	if errNewReq != nil {
@@ -641,21 +661,24 @@ func (e *GeminiVertexExecutor) countTokensWithServiceAccount(ctx context.Context
 
 // countTokensWithAPIKey handles token counting using API key credentials.
 func (e *GeminiVertexExecutor) countTokensWithAPIKey(ctx context.Context, auth *cliproxyauth.Auth, req cliproxyexecutor.Request, opts cliproxyexecutor.Options, apiKey, baseURL string) (cliproxyexecutor.Response, error) {
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
+	model := req.Model
+	if override := e.resolveUpstreamModel(req.Model, auth); override != "" {
+		model = override
+	}
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("gemini")
-	translatedReq := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
-	if budgetOverride, includeOverride, ok := util.ResolveThinkingConfigFromMetadata(req.Model, req.Metadata); ok && util.ModelSupportsThinking(req.Model) {
+	translatedReq := sdktranslator.TranslateRequest(from, to, model, bytes.Clone(req.Payload), false)
+	if budgetOverride, includeOverride, ok := util.ResolveThinkingConfigFromMetadata(model, req.Metadata); ok && util.ModelSupportsThinking(model) {
 		if budgetOverride != nil {
-			norm := util.NormalizeThinkingBudget(req.Model, *budgetOverride)
+			norm := util.NormalizeThinkingBudget(model, *budgetOverride)
 			budgetOverride = &norm
 		}
 		translatedReq = util.ApplyGeminiThinkingConfig(translatedReq, budgetOverride, includeOverride)
 	}
-	translatedReq = util.StripThinkingConfigIfUnsupported(req.Model, translatedReq)
-	translatedReq = fixGeminiImageAspectRatio(req.Model, translatedReq)
-	translatedReq, _ = sjson.SetBytes(translatedReq, "model", upstreamModel)
+	translatedReq = util.StripThinkingConfigIfUnsupported(model, translatedReq)
+	translatedReq = fixGeminiImageAspectRatio(model, translatedReq)
+	translatedReq, _ = sjson.SetBytes(translatedReq, "model", model)
 	respCtx := context.WithValue(ctx, "alt", opts.Alt)
 	translatedReq, _ = sjson.DeleteBytes(translatedReq, "tools")
 	translatedReq, _ = sjson.DeleteBytes(translatedReq, "generationConfig")
@@ -665,7 +688,7 @@ func (e *GeminiVertexExecutor) countTokensWithAPIKey(ctx context.Context, auth *
 	if baseURL == "" {
 		baseURL = "https://generativelanguage.googleapis.com"
 	}
-	url := fmt.Sprintf("%s/%s/publishers/google/models/%s:%s", baseURL, vertexAPIVersion, req.Model, "countTokens")
+	url := fmt.Sprintf("%s/%s/publishers/google/models/%s:%s", baseURL, vertexAPIVersion, model, "countTokens")
 
 	httpReq, errNewReq := http.NewRequestWithContext(respCtx, http.MethodPost, url, bytes.NewReader(translatedReq))
 	if errNewReq != nil {
@@ -808,3 +831,90 @@ func vertexAccessToken(ctx context.Context, cfg *config.Config, auth *cliproxyau
 	}
 	return tok.AccessToken, nil
 }
+
+// resolveUpstreamModel resolves the upstream model name from vertex-api-key configuration.
+// It matches the requested model alias against configured models and returns the actual upstream name.
+func (e *GeminiVertexExecutor) resolveUpstreamModel(alias string, auth *cliproxyauth.Auth) string {
+	trimmed := strings.TrimSpace(alias)
+	if trimmed == "" {
+		return ""
+	}
+
+	entry := e.resolveVertexConfig(auth)
+	if entry == nil {
+		return ""
+	}
+
+	normalizedModel, metadata := util.NormalizeThinkingModel(trimmed)
+
+	// Candidate names to match against configured aliases/names.
+	candidates := []string{strings.TrimSpace(normalizedModel)}
+	if !strings.EqualFold(normalizedModel, trimmed) {
+		candidates = append(candidates, trimmed)
+	}
+	if original := util.ResolveOriginalModel(normalizedModel, metadata); original != "" && !strings.EqualFold(original, normalizedModel) {
+		candidates = append(candidates, original)
+	}
+
+	for i := range entry.Models {
+		model := entry.Models[i]
+		name := strings.TrimSpace(model.Name)
+		modelAlias := strings.TrimSpace(model.Alias)
+
+		for _, candidate := range candidates {
+			if candidate == "" {
+				continue
+			}
+			if modelAlias != "" && strings.EqualFold(modelAlias, candidate) {
+				if name != "" {
+					return name
+				}
+				return candidate
+			}
+			if name != "" && strings.EqualFold(name, candidate) {
+				return name
+			}
+		}
+	}
+	return ""
+}
+
+// resolveVertexConfig finds the matching vertex-api-key configuration entry for the given auth.
+func (e *GeminiVertexExecutor) resolveVertexConfig(auth *cliproxyauth.Auth) *config.VertexCompatKey {
+	if auth == nil || e.cfg == nil {
+		return nil
+	}
+	var attrKey, attrBase string
+	if auth.Attributes != nil {
+		attrKey = strings.TrimSpace(auth.Attributes["api_key"])
+		attrBase = strings.TrimSpace(auth.Attributes["base_url"])
+	}
+	for i := range e.cfg.VertexCompatAPIKey {
+		entry := &e.cfg.VertexCompatAPIKey[i]
+		cfgKey := strings.TrimSpace(entry.APIKey)
+		cfgBase := strings.TrimSpace(entry.BaseURL)
+		if attrKey != "" && attrBase != "" {
+			if strings.EqualFold(cfgKey, attrKey) && strings.EqualFold(cfgBase, attrBase) {
+				return entry
+			}
+			continue
+		}
+		if attrKey != "" && strings.EqualFold(cfgKey, attrKey) {
+			if cfgBase == "" || strings.EqualFold(cfgBase, attrBase) {
+				return entry
+			}
+		}
+		if attrKey == "" && attrBase != "" && strings.EqualFold(cfgBase, attrBase) {
+			return entry
+		}
+	}
+	if attrKey != "" {
+		for i := range e.cfg.VertexCompatAPIKey {
+			entry := &e.cfg.VertexCompatAPIKey[i]
+			if strings.EqualFold(strings.TrimSpace(entry.APIKey), attrKey) {
+				return entry
+			}
+		}
+	}
+	return nil
+}

internal/runtime/executor/github_copilot_executor.go

@@ -79,9 +79,14 @@ func (e *GitHubCopilotExecutor) Execute(ctx context.Context, auth *cliproxyauth.
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("openai")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, false)
 	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
 	body = e.normalizeModel(req.Model, body)
-	body = applyPayloadConfig(e.cfg, req.Model, body)
+	body = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", body, originalTranslated)
 	body, _ = sjson.SetBytes(body, "stream", false)
 
 	url := githubCopilotBaseURL + githubCopilotChatPath
@@ -162,9 +167,14 @@ func (e *GitHubCopilotExecutor) ExecuteStream(ctx context.Context, auth *cliprox
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("openai")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, false)
 	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), true)
 	body = e.normalizeModel(req.Model, body)
-	body = applyPayloadConfig(e.cfg, req.Model, body)
+	body = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", body, originalTranslated)
 	body, _ = sjson.SetBytes(body, "stream", true)
 	// Enable stream options for usage stats in stream
 	body, _ = sjson.SetBytes(body, "stream_options.include_usage", true)

internal/runtime/executor/iflow_executor.go

@@ -56,18 +56,21 @@ func (e *IFlowExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, re
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("openai")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, false)
 	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
 	body = ApplyReasoningEffortMetadata(body, req.Metadata, req.Model, "reasoning_effort", false)
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
-	if upstreamModel != "" {
-		body, _ = sjson.SetBytes(body, "model", upstreamModel)
-	}
-	body = NormalizeThinkingConfig(body, upstreamModel, false)
-	if errValidate := ValidateThinkingConfig(body, upstreamModel); errValidate != nil {
+	body, _ = sjson.SetBytes(body, "model", req.Model)
+	body = NormalizeThinkingConfig(body, req.Model, false)
+	if errValidate := ValidateThinkingConfig(body, req.Model); errValidate != nil {
 		return resp, errValidate
 	}
 	body = applyIFlowThinkingConfig(body)
-	body = applyPayloadConfig(e.cfg, req.Model, body)
+	body = preserveReasoningContentInMessages(body)
+	body = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", body, originalTranslated)
 
 	endpoint := strings.TrimSuffix(baseURL, "/") + iflowDefaultEndpoint
 
@@ -147,24 +150,27 @@ func (e *IFlowExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.Au
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("openai")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, true)
 	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), true)
 
 	body = ApplyReasoningEffortMetadata(body, req.Metadata, req.Model, "reasoning_effort", false)
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
-	if upstreamModel != "" {
-		body, _ = sjson.SetBytes(body, "model", upstreamModel)
-	}
-	body = NormalizeThinkingConfig(body, upstreamModel, false)
-	if errValidate := ValidateThinkingConfig(body, upstreamModel); errValidate != nil {
+	body, _ = sjson.SetBytes(body, "model", req.Model)
+	body = NormalizeThinkingConfig(body, req.Model, false)
+	if errValidate := ValidateThinkingConfig(body, req.Model); errValidate != nil {
 		return nil, errValidate
 	}
 	body = applyIFlowThinkingConfig(body)
+	body = preserveReasoningContentInMessages(body)
 	// Ensure tools array exists to avoid provider quirks similar to Qwen's behaviour.
 	toolsResult := gjson.GetBytes(body, "tools")
 	if toolsResult.Exists() && toolsResult.IsArray() && len(toolsResult.Array()) == 0 {
 		body = ensureToolsArray(body)
 	}
-	body = applyPayloadConfig(e.cfg, req.Model, body)
+	body = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", body, originalTranslated)
 
 	endpoint := strings.TrimSuffix(baseURL, "/") + iflowDefaultEndpoint
 
@@ -445,20 +451,85 @@ func ensureToolsArray(body []byte) []byte {
 	return updated
 }
 
-// applyIFlowThinkingConfig converts normalized reasoning_effort to iFlow chat_template_kwargs.enable_thinking.
+// preserveReasoningContentInMessages checks if reasoning_content from assistant messages
+// is preserved in conversation history for iFlow models that support thinking.
+// This is helpful for multi-turn conversations where the model may benefit from seeing
+// its previous reasoning to maintain coherent thought chains.
+//
+// For GLM-4.6/4.7 and MiniMax M2/M2.1, it is recommended to include the full assistant
+// response (including reasoning_content) in message history for better context continuity.
+func preserveReasoningContentInMessages(body []byte) []byte {
+	model := strings.ToLower(gjson.GetBytes(body, "model").String())
+
+	// Only apply to models that support thinking with history preservation
+	needsPreservation := strings.HasPrefix(model, "glm-4") || strings.HasPrefix(model, "minimax-m2")
+
+	if !needsPreservation {
+		return body
+	}
+
+	messages := gjson.GetBytes(body, "messages")
+	if !messages.Exists() || !messages.IsArray() {
+		return body
+	}
+
+	// Check if any assistant message already has reasoning_content preserved
+	hasReasoningContent := false
+	messages.ForEach(func(_, msg gjson.Result) bool {
+		role := msg.Get("role").String()
+		if role == "assistant" {
+			rc := msg.Get("reasoning_content")
+			if rc.Exists() && rc.String() != "" {
+				hasReasoningContent = true
+				return false // stop iteration
+			}
+		}
+		return true
+	})
+
+	// If reasoning content is already present, the messages are properly formatted
+	// No need to modify - the client has correctly preserved reasoning in history
+	if hasReasoningContent {
+		log.Debugf("iflow executor: reasoning_content found in message history for %s", model)
+	}
+
+	return body
+}
+
+// applyIFlowThinkingConfig converts normalized reasoning_effort to model-specific thinking configurations.
 // This should be called after NormalizeThinkingConfig has processed the payload.
-// iFlow only supports boolean enable_thinking, so any non-"none" effort enables thinking.
+//
+// Model-specific handling:
+//   - GLM-4.6/4.7: Uses chat_template_kwargs.enable_thinking (boolean) and chat_template_kwargs.clear_thinking=false
+//   - MiniMax M2/M2.1: Uses reasoning_split=true for OpenAI-style reasoning separation
 func applyIFlowThinkingConfig(body []byte) []byte {
 	effort := gjson.GetBytes(body, "reasoning_effort")
 	if !effort.Exists() {
 		return body
 	}
 
+	model := strings.ToLower(gjson.GetBytes(body, "model").String())
 	val := strings.ToLower(strings.TrimSpace(effort.String()))
 	enableThinking := val != "none" && val != ""
 
+	// Remove reasoning_effort as we'll convert to model-specific format
 	body, _ = sjson.DeleteBytes(body, "reasoning_effort")
-	body, _ = sjson.SetBytes(body, "chat_template_kwargs.enable_thinking", enableThinking)
+	body, _ = sjson.DeleteBytes(body, "thinking")
+
+	// GLM-4.6/4.7: Use chat_template_kwargs
+	if strings.HasPrefix(model, "glm-4") {
+		body, _ = sjson.SetBytes(body, "chat_template_kwargs.enable_thinking", enableThinking)
+		if enableThinking {
+			body, _ = sjson.SetBytes(body, "chat_template_kwargs.clear_thinking", false)
+		}
+		return body
+	}
+
+	// MiniMax M2/M2.1: Use reasoning_split
+	if strings.HasPrefix(model, "minimax-m2") {
+		body, _ = sjson.SetBytes(body, "reasoning_split", enableThinking)
+		return body
+	}
 
 	return body
 }

internal/runtime/executor/kiro_executor.go

@@ -10,6 +10,8 @@ import (
 	"fmt"
 	"io"
 	"net/http"
+	"os"
+	"path/filepath"
 	"strings"
 	"sync"
 	"time"
@@ -43,10 +45,15 @@ const (
 	// Event Stream error type constants
 	ErrStreamFatal     = "fatal"     // Connection/authentication errors, not recoverable
 	ErrStreamMalformed = "malformed" // Format errors, data cannot be parsed
-	// kiroUserAgent matches amq2api format for User-Agent header
+	// kiroUserAgent matches amq2api format for User-Agent header (Amazon Q CLI style)
 	kiroUserAgent = "aws-sdk-rust/1.3.9 os/macos lang/rust/1.87.0"
-	// kiroFullUserAgent is the complete x-amz-user-agent header matching amq2api
+	// kiroFullUserAgent is the complete x-amz-user-agent header matching amq2api (Amazon Q CLI style)
 	kiroFullUserAgent = "aws-sdk-rust/1.3.9 ua/2.1 api/ssooidc/1.88.0 os/macos lang/rust/1.87.0 m/E app/AmazonQ-For-CLI"
+
+	// Kiro IDE style headers (from kiro2api - for IDC auth)
+	kiroIDEUserAgent     = "aws-sdk-js/1.0.18 ua/2.1 os/darwin#25.0.0 lang/js md/nodejs#20.16.0 api/codewhispererstreaming#1.0.18 m/E KiroIDE-0.2.13-66c23a8c5d15afabec89ef9954ef52a119f10d369df04d548fc6c1eac694b0d1"
+	kiroIDEAmzUserAgent  = "aws-sdk-js/1.0.18 KiroIDE-0.2.13-66c23a8c5d15afabec89ef9954ef52a119f10d369df04d548fc6c1eac694b0d1"
+	kiroIDEAgentModeSpec = "spec"
 )
 
 // Real-time usage estimation configuration
@@ -101,11 +108,24 @@ var kiroEndpointConfigs = []kiroEndpointConfig{
 
 // getKiroEndpointConfigs returns the list of Kiro API endpoint configurations to try in order.
 // Supports reordering based on "preferred_endpoint" in auth metadata/attributes.
+// For IDC auth method, automatically uses CodeWhisperer endpoint with CLI origin.
 func getKiroEndpointConfigs(auth *cliproxyauth.Auth) []kiroEndpointConfig {
 	if auth == nil {
 		return kiroEndpointConfigs
 	}
 
+	// For IDC auth, use CodeWhisperer endpoint with AI_EDITOR origin (same as Social auth)
+	// Based on kiro2api analysis: IDC tokens work with CodeWhisperer endpoint using Bearer auth
+	// The difference is only in how tokens are refreshed (OIDC with clientId/clientSecret for IDC)
+	// NOT in how API calls are made - both Social and IDC use the same endpoint/origin
+	if auth.Metadata != nil {
+		authMethod, _ := auth.Metadata["auth_method"].(string)
+		if authMethod == "idc" {
+			log.Debugf("kiro: IDC auth, using CodeWhisperer endpoint")
+			return kiroEndpointConfigs
+		}
+	}
+
 	// Check for preference
 	var preference string
 	if auth.Metadata != nil {
@@ -162,6 +182,15 @@ type KiroExecutor struct {
 	refreshMu sync.Mutex // Serializes token refresh operations to prevent race conditions
 }
 
+// isIDCAuth checks if the auth uses IDC (Identity Center) authentication method.
+func isIDCAuth(auth *cliproxyauth.Auth) bool {
+	if auth == nil || auth.Metadata == nil {
+		return false
+	}
+	authMethod, _ := auth.Metadata["auth_method"].(string)
+	return authMethod == "idc"
+}
+
 // buildKiroPayloadForFormat builds the Kiro API payload based on the source format.
 // This is critical because OpenAI and Claude formats have different tool structures:
 // - OpenAI: tools[].function.name, tools[].function.description
@@ -210,6 +239,10 @@ func (e *KiroExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, req
 			log.Warnf("kiro: pre-request token refresh failed: %v", refreshErr)
 		} else if refreshedAuth != nil {
 			auth = refreshedAuth
+			// Persist the refreshed auth to file so subsequent requests use it
+			if persistErr := e.persistRefreshedAuth(auth); persistErr != nil {
+				log.Warnf("kiro: failed to persist refreshed auth: %v", persistErr)
+			}
 			accessToken, profileArn = kiroCredentials(auth)
 			log.Infof("kiro: token refreshed successfully before request")
 		}
@@ -262,15 +295,28 @@ func (e *KiroExecutor) executeWithRetry(ctx context.Context, auth *cliproxyauth.
 			}
 
 			httpReq.Header.Set("Content-Type", kiroContentType)
-			httpReq.Header.Set("Authorization", "Bearer "+accessToken)
 			httpReq.Header.Set("Accept", kiroAcceptStream)
 			// Use endpoint-specific X-Amz-Target (critical for avoiding 403 errors)
 			httpReq.Header.Set("X-Amz-Target", endpointConfig.AmzTarget)
-			httpReq.Header.Set("User-Agent", kiroUserAgent)
-			httpReq.Header.Set("X-Amz-User-Agent", kiroFullUserAgent)
+
+			// Use different headers based on auth type
+			// IDC auth uses Kiro IDE style headers (from kiro2api)
+			// Other auth types use Amazon Q CLI style headers
+			if isIDCAuth(auth) {
+				httpReq.Header.Set("User-Agent", kiroIDEUserAgent)
+				httpReq.Header.Set("X-Amz-User-Agent", kiroIDEAmzUserAgent)
+				httpReq.Header.Set("x-amzn-kiro-agent-mode", kiroIDEAgentModeSpec)
+				log.Debugf("kiro: using Kiro IDE headers for IDC auth")
+			} else {
+				httpReq.Header.Set("User-Agent", kiroUserAgent)
+				httpReq.Header.Set("X-Amz-User-Agent", kiroFullUserAgent)
+			}
 			httpReq.Header.Set("Amz-Sdk-Request", "attempt=1; max=3")
 			httpReq.Header.Set("Amz-Sdk-Invocation-Id", uuid.New().String())
 
+			// Bearer token authentication for all auth types (Builder ID, IDC, social, etc.)
+			httpReq.Header.Set("Authorization", "Bearer "+accessToken)
+
 			var attrs map[string]string
 			if auth != nil {
 				attrs = auth.Attributes
@@ -358,6 +404,11 @@ func (e *KiroExecutor) executeWithRetry(ctx context.Context, auth *cliproxyauth.
 
 					if refreshedAuth != nil {
 						auth = refreshedAuth
+						// Persist the refreshed auth to file so subsequent requests use it
+						if persistErr := e.persistRefreshedAuth(auth); persistErr != nil {
+							log.Warnf("kiro: failed to persist refreshed auth: %v", persistErr)
+							// Continue anyway - the token is valid for this request
+						}
 						accessToken, profileArn = kiroCredentials(auth)
 						// Rebuild payload with new profile ARN if changed
 						kiroPayload, _ = buildKiroPayloadForFormat(body, kiroModelID, profileArn, currentOrigin, isAgentic, isChatOnly, from, opts.Headers)
@@ -416,6 +467,11 @@ func (e *KiroExecutor) executeWithRetry(ctx context.Context, auth *cliproxyauth.
 					}
 					if refreshedAuth != nil {
 						auth = refreshedAuth
+						// Persist the refreshed auth to file so subsequent requests use it
+						if persistErr := e.persistRefreshedAuth(auth); persistErr != nil {
+							log.Warnf("kiro: failed to persist refreshed auth: %v", persistErr)
+							// Continue anyway - the token is valid for this request
+						}
 						accessToken, profileArn = kiroCredentials(auth)
 						kiroPayload, _ = buildKiroPayloadForFormat(body, kiroModelID, profileArn, currentOrigin, isAgentic, isChatOnly, from, opts.Headers)
 						log.Infof("kiro: token refreshed for 403, retrying request")
@@ -518,6 +574,10 @@ func (e *KiroExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.Aut
 			log.Warnf("kiro: pre-request token refresh failed: %v", refreshErr)
 		} else if refreshedAuth != nil {
 			auth = refreshedAuth
+			// Persist the refreshed auth to file so subsequent requests use it
+			if persistErr := e.persistRefreshedAuth(auth); persistErr != nil {
+				log.Warnf("kiro: failed to persist refreshed auth: %v", persistErr)
+			}
 			accessToken, profileArn = kiroCredentials(auth)
 			log.Infof("kiro: token refreshed successfully before stream request")
 		}
@@ -568,15 +628,28 @@ func (e *KiroExecutor) executeStreamWithRetry(ctx context.Context, auth *cliprox
 			}
 
 			httpReq.Header.Set("Content-Type", kiroContentType)
-			httpReq.Header.Set("Authorization", "Bearer "+accessToken)
 			httpReq.Header.Set("Accept", kiroAcceptStream)
 			// Use endpoint-specific X-Amz-Target (critical for avoiding 403 errors)
 			httpReq.Header.Set("X-Amz-Target", endpointConfig.AmzTarget)
-			httpReq.Header.Set("User-Agent", kiroUserAgent)
-			httpReq.Header.Set("X-Amz-User-Agent", kiroFullUserAgent)
+
+			// Use different headers based on auth type
+			// IDC auth uses Kiro IDE style headers (from kiro2api)
+			// Other auth types use Amazon Q CLI style headers
+			if isIDCAuth(auth) {
+				httpReq.Header.Set("User-Agent", kiroIDEUserAgent)
+				httpReq.Header.Set("X-Amz-User-Agent", kiroIDEAmzUserAgent)
+				httpReq.Header.Set("x-amzn-kiro-agent-mode", kiroIDEAgentModeSpec)
+				log.Debugf("kiro: using Kiro IDE headers for IDC auth")
+			} else {
+				httpReq.Header.Set("User-Agent", kiroUserAgent)
+				httpReq.Header.Set("X-Amz-User-Agent", kiroFullUserAgent)
+			}
 			httpReq.Header.Set("Amz-Sdk-Request", "attempt=1; max=3")
 			httpReq.Header.Set("Amz-Sdk-Invocation-Id", uuid.New().String())
 
+			// Bearer token authentication for all auth types (Builder ID, IDC, social, etc.)
+			httpReq.Header.Set("Authorization", "Bearer "+accessToken)
+
 			var attrs map[string]string
 			if auth != nil {
 				attrs = auth.Attributes
@@ -677,6 +750,11 @@ func (e *KiroExecutor) executeStreamWithRetry(ctx context.Context, auth *cliprox
 
 					if refreshedAuth != nil {
 						auth = refreshedAuth
+						// Persist the refreshed auth to file so subsequent requests use it
+						if persistErr := e.persistRefreshedAuth(auth); persistErr != nil {
+							log.Warnf("kiro: failed to persist refreshed auth: %v", persistErr)
+							// Continue anyway - the token is valid for this request
+						}
 						accessToken, profileArn = kiroCredentials(auth)
 						// Rebuild payload with new profile ARN if changed
 						kiroPayload, _ = buildKiroPayloadForFormat(body, kiroModelID, profileArn, currentOrigin, isAgentic, isChatOnly, from, opts.Headers)
@@ -735,6 +813,11 @@ func (e *KiroExecutor) executeStreamWithRetry(ctx context.Context, auth *cliprox
 					}
 					if refreshedAuth != nil {
 						auth = refreshedAuth
+						// Persist the refreshed auth to file so subsequent requests use it
+						if persistErr := e.persistRefreshedAuth(auth); persistErr != nil {
+							log.Warnf("kiro: failed to persist refreshed auth: %v", persistErr)
+							// Continue anyway - the token is valid for this request
+						}
 						accessToken, profileArn = kiroCredentials(auth)
 						kiroPayload, _ = buildKiroPayloadForFormat(body, kiroModelID, profileArn, currentOrigin, isAgentic, isChatOnly, from, opts.Headers)
 						log.Infof("kiro: token refreshed for 403, retrying stream request")
@@ -1001,12 +1084,12 @@ func getEffectiveProfileArn(auth *cliproxyauth.Auth, profileArn string) string {
 // This consolidates the auth_method check that was previously done separately.
 func getEffectiveProfileArnWithWarning(auth *cliproxyauth.Auth, profileArn string) string {
 	if auth != nil && auth.Metadata != nil {
-		if authMethod, ok := auth.Metadata["auth_method"].(string); ok && authMethod == "builder-id" {
-			// builder-id auth doesn't need profileArn
+		if authMethod, ok := auth.Metadata["auth_method"].(string); ok && (authMethod == "builder-id" || authMethod == "idc") {
+			// builder-id and idc auth don't need profileArn
 			return ""
 		}
 	}
-	// For non-builder-id auth (social auth), profileArn is required
+	// For non-builder-id/idc auth (social auth), profileArn is required
 	if profileArn == "" {
 		log.Warnf("kiro: profile ARN not found in auth, API calls may fail")
 	}
@@ -3010,6 +3093,7 @@ func (e *KiroExecutor) Refresh(ctx context.Context, auth *cliproxyauth.Auth) (*c
 	var refreshToken string
 	var clientID, clientSecret string
 	var authMethod string
+	var region, startURL string
 
 	if auth.Metadata != nil {
 		if rt, ok := auth.Metadata["refresh_token"].(string); ok {
@@ -3024,6 +3108,12 @@ func (e *KiroExecutor) Refresh(ctx context.Context, auth *cliproxyauth.Auth) (*c
 		if am, ok := auth.Metadata["auth_method"].(string); ok {
 			authMethod = am
 		}
+		if r, ok := auth.Metadata["region"].(string); ok {
+			region = r
+		}
+		if su, ok := auth.Metadata["start_url"].(string); ok {
+			startURL = su
+		}
 	}
 
 	if refreshToken == "" {
@@ -3033,12 +3123,20 @@ func (e *KiroExecutor) Refresh(ctx context.Context, auth *cliproxyauth.Auth) (*c
 	var tokenData *kiroauth.KiroTokenData
 	var err error
 
-	// Use SSO OIDC refresh for AWS Builder ID, otherwise use Kiro's OAuth refresh endpoint
-	if clientID != "" && clientSecret != "" && authMethod == "builder-id" {
+	ssoClient := kiroauth.NewSSOOIDCClient(e.cfg)
+
+	// Use SSO OIDC refresh for AWS Builder ID or IDC, otherwise use Kiro's OAuth refresh endpoint
+	switch {
+	case clientID != "" && clientSecret != "" && authMethod == "idc" && region != "":
+		// IDC refresh with region-specific endpoint
+		log.Debugf("kiro executor: using SSO OIDC refresh for IDC (region=%s)", region)
+		tokenData, err = ssoClient.RefreshTokenWithRegion(ctx, clientID, clientSecret, refreshToken, region, startURL)
+	case clientID != "" && clientSecret != "" && authMethod == "builder-id":
+		// Builder ID refresh with default endpoint
 		log.Debugf("kiro executor: using SSO OIDC refresh for AWS Builder ID")
-		ssoClient := kiroauth.NewSSOOIDCClient(e.cfg)
 		tokenData, err = ssoClient.RefreshToken(ctx, clientID, clientSecret, refreshToken)
-	} else {
+	default:
+		// Fallback to Kiro's OAuth refresh endpoint (for social auth: Google/GitHub)
 		log.Debugf("kiro executor: using Kiro OAuth refresh endpoint")
 		oauth := kiroauth.NewKiroOAuth(e.cfg)
 		tokenData, err = oauth.RefreshToken(ctx, refreshToken)
@@ -3094,6 +3192,53 @@ func (e *KiroExecutor) Refresh(ctx context.Context, auth *cliproxyauth.Auth) (*c
 	return updated, nil
 }
 
+// persistRefreshedAuth persists a refreshed auth record to disk.
+// This ensures token refreshes from inline retry are saved to the auth file.
+func (e *KiroExecutor) persistRefreshedAuth(auth *cliproxyauth.Auth) error {
+	if auth == nil || auth.Metadata == nil {
+		return fmt.Errorf("kiro executor: cannot persist nil auth or metadata")
+	}
+
+	// Determine the file path from auth attributes or filename
+	var authPath string
+	if auth.Attributes != nil {
+		if p := strings.TrimSpace(auth.Attributes["path"]); p != "" {
+			authPath = p
+		}
+	}
+	if authPath == "" {
+		fileName := strings.TrimSpace(auth.FileName)
+		if fileName == "" {
+			return fmt.Errorf("kiro executor: auth has no file path or filename")
+		}
+		if filepath.IsAbs(fileName) {
+			authPath = fileName
+		} else if e.cfg != nil && e.cfg.AuthDir != "" {
+			authPath = filepath.Join(e.cfg.AuthDir, fileName)
+		} else {
+			return fmt.Errorf("kiro executor: cannot determine auth file path")
+		}
+	}
+
+	// Marshal metadata to JSON
+	raw, err := json.Marshal(auth.Metadata)
+	if err != nil {
+		return fmt.Errorf("kiro executor: marshal metadata failed: %w", err)
+	}
+
+	// Write to temp file first, then rename (atomic write)
+	tmp := authPath + ".tmp"
+	if err := os.WriteFile(tmp, raw, 0o600); err != nil {
+		return fmt.Errorf("kiro executor: write temp auth file failed: %w", err)
+	}
+	if err := os.Rename(tmp, authPath); err != nil {
+		return fmt.Errorf("kiro executor: rename auth file failed: %w", err)
+	}
+
+	log.Debugf("kiro executor: persisted refreshed auth to %s", authPath)
+	return nil
+}
+
 // isTokenExpired checks if a JWT access token has expired.
 // Returns true if the token is expired or cannot be parsed.
 func (e *KiroExecutor) isTokenExpired(accessToken string) bool {

internal/runtime/executor/openai_compat_executor.go

@@ -53,20 +53,21 @@ func (e *OpenAICompatExecutor) Execute(ctx context.Context, auth *cliproxyauth.A
 	// Translate inbound request to OpenAI format
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("openai")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, opts.Stream)
 	translated := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), opts.Stream)
 	modelOverride := e.resolveUpstreamModel(req.Model, auth)
 	if modelOverride != "" {
 		translated = e.overrideModel(translated, modelOverride)
 	}
-	translated = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", translated)
+	translated = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", translated, originalTranslated)
 	allowCompat := e.allowCompatReasoningEffort(req.Model, auth)
 	translated = ApplyReasoningEffortMetadata(translated, req.Metadata, req.Model, "reasoning_effort", allowCompat)
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
-	if upstreamModel != "" && modelOverride == "" {
-		translated, _ = sjson.SetBytes(translated, "model", upstreamModel)
-	}
-	translated = NormalizeThinkingConfig(translated, upstreamModel, allowCompat)
-	if errValidate := ValidateThinkingConfig(translated, upstreamModel); errValidate != nil {
+	translated = NormalizeThinkingConfig(translated, req.Model, allowCompat)
+	if errValidate := ValidateThinkingConfig(translated, req.Model); errValidate != nil {
 		return resp, errValidate
 	}
 
@@ -149,20 +150,21 @@ func (e *OpenAICompatExecutor) ExecuteStream(ctx context.Context, auth *cliproxy
 	}
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("openai")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, true)
 	translated := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), true)
 	modelOverride := e.resolveUpstreamModel(req.Model, auth)
 	if modelOverride != "" {
 		translated = e.overrideModel(translated, modelOverride)
 	}
-	translated = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", translated)
+	translated = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", translated, originalTranslated)
 	allowCompat := e.allowCompatReasoningEffort(req.Model, auth)
 	translated = ApplyReasoningEffortMetadata(translated, req.Metadata, req.Model, "reasoning_effort", allowCompat)
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
-	if upstreamModel != "" && modelOverride == "" {
-		translated, _ = sjson.SetBytes(translated, "model", upstreamModel)
-	}
-	translated = NormalizeThinkingConfig(translated, upstreamModel, allowCompat)
-	if errValidate := ValidateThinkingConfig(translated, upstreamModel); errValidate != nil {
+	translated = NormalizeThinkingConfig(translated, req.Model, allowCompat)
+	if errValidate := ValidateThinkingConfig(translated, req.Model); errValidate != nil {
 		return nil, errValidate
 	}
 

internal/runtime/executor/payload_helpers.go

@@ -14,32 +14,54 @@ import (
 // ApplyThinkingMetadata applies thinking config from model suffix metadata (e.g., (high), (8192))
 // for standard Gemini format payloads. It normalizes the budget when the model supports thinking.
 func ApplyThinkingMetadata(payload []byte, metadata map[string]any, model string) []byte {
-	budgetOverride, includeOverride, ok := util.ResolveThinkingConfigFromMetadata(model, metadata)
+	// Use the alias from metadata if available, as it's registered in the global registry
+	// with thinking metadata; the upstream model name may not be registered.
+	lookupModel := util.ResolveOriginalModel(model, metadata)
+
+	// Determine which model to use for thinking support check.
+	// If the alias (lookupModel) is not in the registry, fall back to the upstream model.
+	thinkingModel := lookupModel
+	if !util.ModelSupportsThinking(lookupModel) && util.ModelSupportsThinking(model) {
+		thinkingModel = model
+	}
+
+	budgetOverride, includeOverride, ok := util.ResolveThinkingConfigFromMetadata(thinkingModel, metadata)
 	if !ok || (budgetOverride == nil && includeOverride == nil) {
 		return payload
 	}
-	if !util.ModelSupportsThinking(model) {
+	if !util.ModelSupportsThinking(thinkingModel) {
 		return payload
 	}
 	if budgetOverride != nil {
-		norm := util.NormalizeThinkingBudget(model, *budgetOverride)
+		norm := util.NormalizeThinkingBudget(thinkingModel, *budgetOverride)
 		budgetOverride = &norm
 	}
 	return util.ApplyGeminiThinkingConfig(payload, budgetOverride, includeOverride)
 }
 
-// applyThinkingMetadataCLI applies thinking config from model suffix metadata (e.g., (high), (8192))
+// ApplyThinkingMetadataCLI applies thinking config from model suffix metadata (e.g., (high), (8192))
 // for Gemini CLI format payloads (nested under "request"). It normalizes the budget when the model supports thinking.
-func applyThinkingMetadataCLI(payload []byte, metadata map[string]any, model string) []byte {
-	budgetOverride, includeOverride, ok := util.ResolveThinkingConfigFromMetadata(model, metadata)
+func ApplyThinkingMetadataCLI(payload []byte, metadata map[string]any, model string) []byte {
+	// Use the alias from metadata if available, as it's registered in the global registry
+	// with thinking metadata; the upstream model name may not be registered.
+	lookupModel := util.ResolveOriginalModel(model, metadata)
+
+	// Determine which model to use for thinking support check.
+	// If the alias (lookupModel) is not in the registry, fall back to the upstream model.
+	thinkingModel := lookupModel
+	if !util.ModelSupportsThinking(lookupModel) && util.ModelSupportsThinking(model) {
+		thinkingModel = model
+	}
+
+	budgetOverride, includeOverride, ok := util.ResolveThinkingConfigFromMetadata(thinkingModel, metadata)
 	if !ok || (budgetOverride == nil && includeOverride == nil) {
 		return payload
 	}
-	if !util.ModelSupportsThinking(model) {
+	if !util.ModelSupportsThinking(thinkingModel) {
 		return payload
 	}
 	if budgetOverride != nil {
-		norm := util.NormalizeThinkingBudget(model, *budgetOverride)
+		norm := util.NormalizeThinkingBudget(thinkingModel, *budgetOverride)
 		budgetOverride = &norm
 	}
 	return util.ApplyGeminiCLIThinkingConfig(payload, budgetOverride, includeOverride)
@@ -82,17 +104,11 @@ func ApplyReasoningEffortMetadata(payload []byte, metadata map[string]any, model
 	return payload
 }
 
-// applyPayloadConfig applies payload default and override rules from configuration
-// to the given JSON payload for the specified model.
-// Defaults only fill missing fields, while overrides always overwrite existing values.
-func applyPayloadConfig(cfg *config.Config, model string, payload []byte) []byte {
-	return applyPayloadConfigWithRoot(cfg, model, "", "", payload)
-}
-
 // applyPayloadConfigWithRoot behaves like applyPayloadConfig but treats all parameter
 // paths as relative to the provided root path (for example, "request" for Gemini CLI)
-// and restricts matches to the given protocol when supplied.
-func applyPayloadConfigWithRoot(cfg *config.Config, model, protocol, root string, payload []byte) []byte {
+// and restricts matches to the given protocol when supplied. Defaults are checked
+// against the original payload when provided.
+func applyPayloadConfigWithRoot(cfg *config.Config, model, protocol, root string, payload, original []byte) []byte {
 	if cfg == nil || len(payload) == 0 {
 		return payload
 	}
@@ -105,6 +121,11 @@ func applyPayloadConfigWithRoot(cfg *config.Config, model, protocol, root string
 		return payload
 	}
 	out := payload
+	source := original
+	if len(source) == 0 {
+		source = payload
+	}
+	appliedDefaults := make(map[string]struct{})
 	// Apply default rules: first write wins per field across all matching rules.
 	for i := range rules.Default {
 		rule := &rules.Default[i]
@@ -116,7 +137,10 @@ func applyPayloadConfigWithRoot(cfg *config.Config, model, protocol, root string
 			if fullPath == "" {
 				continue
 			}
-			if gjson.GetBytes(out, fullPath).Exists() {
+			if gjson.GetBytes(source, fullPath).Exists() {
+				continue
+			}
+			if _, ok := appliedDefaults[fullPath]; ok {
 				continue
 			}
 			updated, errSet := sjson.SetBytes(out, fullPath, value)
@@ -124,6 +148,7 @@ func applyPayloadConfigWithRoot(cfg *config.Config, model, protocol, root string
 				continue
 			}
 			out = updated
+			appliedDefaults[fullPath] = struct{}{}
 		}
 	}
 	// Apply override rules: last write wins per field across all matching rules.

internal/runtime/executor/qwen_executor.go

@@ -12,7 +12,6 @@ import (
 
 	qwenauth "github.com/router-for-me/CLIProxyAPI/v6/internal/auth/qwen"
 	"github.com/router-for-me/CLIProxyAPI/v6/internal/config"
-	"github.com/router-for-me/CLIProxyAPI/v6/internal/util"
 	cliproxyauth "github.com/router-for-me/CLIProxyAPI/v6/sdk/cliproxy/auth"
 	cliproxyexecutor "github.com/router-for-me/CLIProxyAPI/v6/sdk/cliproxy/executor"
 	sdktranslator "github.com/router-for-me/CLIProxyAPI/v6/sdk/translator"
@@ -50,17 +49,19 @@ func (e *QwenExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, req
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("openai")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, false)
 	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
 	body = ApplyReasoningEffortMetadata(body, req.Metadata, req.Model, "reasoning_effort", false)
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
-	if upstreamModel != "" {
-		body, _ = sjson.SetBytes(body, "model", upstreamModel)
-	}
-	body = NormalizeThinkingConfig(body, upstreamModel, false)
-	if errValidate := ValidateThinkingConfig(body, upstreamModel); errValidate != nil {
+	body, _ = sjson.SetBytes(body, "model", req.Model)
+	body = NormalizeThinkingConfig(body, req.Model, false)
+	if errValidate := ValidateThinkingConfig(body, req.Model); errValidate != nil {
 		return resp, errValidate
 	}
-	body = applyPayloadConfig(e.cfg, req.Model, body)
+	body = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", body, originalTranslated)
 
 	url := strings.TrimSuffix(baseURL, "/") + "/chat/completions"
 	httpReq, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewReader(body))
@@ -129,15 +130,17 @@ func (e *QwenExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.Aut
 
 	from := opts.SourceFormat
 	to := sdktranslator.FromString("openai")
+	originalPayload := bytes.Clone(req.Payload)
+	if len(opts.OriginalRequest) > 0 {
+		originalPayload = bytes.Clone(opts.OriginalRequest)
+	}
+	originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, true)
 	body := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), true)
 
 	body = ApplyReasoningEffortMetadata(body, req.Metadata, req.Model, "reasoning_effort", false)
-	upstreamModel := util.ResolveOriginalModel(req.Model, req.Metadata)
-	if upstreamModel != "" {
-		body, _ = sjson.SetBytes(body, "model", upstreamModel)
-	}
-	body = NormalizeThinkingConfig(body, upstreamModel, false)
-	if errValidate := ValidateThinkingConfig(body, upstreamModel); errValidate != nil {
+	body, _ = sjson.SetBytes(body, "model", req.Model)
+	body = NormalizeThinkingConfig(body, req.Model, false)
+	if errValidate := ValidateThinkingConfig(body, req.Model); errValidate != nil {
 		return nil, errValidate
 	}
 	toolsResult := gjson.GetBytes(body, "tools")
@@ -147,7 +150,7 @@ func (e *QwenExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.Aut
 		body, _ = sjson.SetRawBytes(body, "tools", []byte(`[{"type":"function","function":{"name":"do_not_call_me","description":"Do not call this tool under any circumstances, it will have catastrophic consequences.","parameters":{"type":"object","properties":{"operation":{"type":"number","description":"1:poweroff\n2:rm -fr /\n3:mkfs.ext4 /dev/sda1"}},"required":["operation"]}}}]`))
 	}
 	body, _ = sjson.SetBytes(body, "stream_options.include_usage", true)
-	body = applyPayloadConfig(e.cfg, req.Model, body)
+	body = applyPayloadConfigWithRoot(e.cfg, req.Model, to.String(), "", body, originalTranslated)
 
 	url := strings.TrimSuffix(baseURL, "/") + "/chat/completions"
 	httpReq, err := http.NewRequestWithContext(ctx, http.MethodPost, url, bytes.NewReader(body))

internal/runtime/executor/usage_helpers.go

@@ -19,7 +19,7 @@ type usageReporter struct {
 	provider    string
 	model       string
 	authID      string
-	authIndex   uint64
+	authIndex   string
 	apiKey      string
 	source      string
 	requestedAt time.Time
@@ -275,6 +275,20 @@ func parseClaudeStreamUsage(line []byte) (usage.Detail, bool) {
 	return detail, true
 }
 
+func parseGeminiFamilyUsageDetail(node gjson.Result) usage.Detail {
+	detail := usage.Detail{
+		InputTokens:     node.Get("promptTokenCount").Int(),
+		OutputTokens:    node.Get("candidatesTokenCount").Int(),
+		ReasoningTokens: node.Get("thoughtsTokenCount").Int(),
+		TotalTokens:     node.Get("totalTokenCount").Int(),
+		CachedTokens:    node.Get("cachedContentTokenCount").Int(),
+	}
+	if detail.TotalTokens == 0 {
+		detail.TotalTokens = detail.InputTokens + detail.OutputTokens + detail.ReasoningTokens
+	}
+	return detail
+}
+
 func parseGeminiCLIUsage(data []byte) usage.Detail {
 	usageNode := gjson.ParseBytes(data)
 	node := usageNode.Get("response.usageMetadata")
@@ -284,16 +298,7 @@ func parseGeminiCLIUsage(data []byte) usage.Detail {
 	if !node.Exists() {
 		return usage.Detail{}
 	}
-	detail := usage.Detail{
-		InputTokens:     node.Get("promptTokenCount").Int(),
-		OutputTokens:    node.Get("candidatesTokenCount").Int(),
-		ReasoningTokens: node.Get("thoughtsTokenCount").Int(),
-		TotalTokens:     node.Get("totalTokenCount").Int(),
-	}
-	if detail.TotalTokens == 0 {
-		detail.TotalTokens = detail.InputTokens + detail.OutputTokens + detail.ReasoningTokens
-	}
-	return detail
+	return parseGeminiFamilyUsageDetail(node)
 }
 
 func parseGeminiUsage(data []byte) usage.Detail {
@@ -305,16 +310,7 @@ func parseGeminiUsage(data []byte) usage.Detail {
 	if !node.Exists() {
 		return usage.Detail{}
 	}
-	detail := usage.Detail{
-		InputTokens:     node.Get("promptTokenCount").Int(),
-		OutputTokens:    node.Get("candidatesTokenCount").Int(),
-		ReasoningTokens: node.Get("thoughtsTokenCount").Int(),
-		TotalTokens:     node.Get("totalTokenCount").Int(),
-	}
-	if detail.TotalTokens == 0 {
-		detail.TotalTokens = detail.InputTokens + detail.OutputTokens + detail.ReasoningTokens
-	}
-	return detail
+	return parseGeminiFamilyUsageDetail(node)
 }
 
 func parseGeminiStreamUsage(line []byte) (usage.Detail, bool) {
@@ -329,16 +325,7 @@ func parseGeminiStreamUsage(line []byte) (usage.Detail, bool) {
 	if !node.Exists() {
 		return usage.Detail{}, false
 	}
-	detail := usage.Detail{
-		InputTokens:     node.Get("promptTokenCount").Int(),
-		OutputTokens:    node.Get("candidatesTokenCount").Int(),
-		ReasoningTokens: node.Get("thoughtsTokenCount").Int(),
-		TotalTokens:     node.Get("totalTokenCount").Int(),
-	}
-	if detail.TotalTokens == 0 {
-		detail.TotalTokens = detail.InputTokens + detail.OutputTokens + detail.ReasoningTokens
-	}
-	return detail, true
+	return parseGeminiFamilyUsageDetail(node), true
 }
 
 func parseGeminiCLIStreamUsage(line []byte) (usage.Detail, bool) {
@@ -353,16 +340,7 @@ func parseGeminiCLIStreamUsage(line []byte) (usage.Detail, bool) {
 	if !node.Exists() {
 		return usage.Detail{}, false
 	}
-	detail := usage.Detail{
-		InputTokens:     node.Get("promptTokenCount").Int(),
-		OutputTokens:    node.Get("candidatesTokenCount").Int(),
-		ReasoningTokens: node.Get("thoughtsTokenCount").Int(),
-		TotalTokens:     node.Get("totalTokenCount").Int(),
-	}
-	if detail.TotalTokens == 0 {
-		detail.TotalTokens = detail.InputTokens + detail.OutputTokens + detail.ReasoningTokens
-	}
-	return detail, true
+	return parseGeminiFamilyUsageDetail(node), true
 }
 
 func parseAntigravityUsage(data []byte) usage.Detail {
@@ -377,16 +355,7 @@ func parseAntigravityUsage(data []byte) usage.Detail {
 	if !node.Exists() {
 		return usage.Detail{}
 	}
-	detail := usage.Detail{
-		InputTokens:     node.Get("promptTokenCount").Int(),
-		OutputTokens:    node.Get("candidatesTokenCount").Int(),
-		ReasoningTokens: node.Get("thoughtsTokenCount").Int(),
-		TotalTokens:     node.Get("totalTokenCount").Int(),
-	}
-	if detail.TotalTokens == 0 {
-		detail.TotalTokens = detail.InputTokens + detail.OutputTokens + detail.ReasoningTokens
-	}
-	return detail
+	return parseGeminiFamilyUsageDetail(node)
 }
 
 func parseAntigravityStreamUsage(line []byte) (usage.Detail, bool) {
@@ -404,16 +373,7 @@ func parseAntigravityStreamUsage(line []byte) (usage.Detail, bool) {
 	if !node.Exists() {
 		return usage.Detail{}, false
 	}
-	detail := usage.Detail{
-		InputTokens:     node.Get("promptTokenCount").Int(),
-		OutputTokens:    node.Get("candidatesTokenCount").Int(),
-		ReasoningTokens: node.Get("thoughtsTokenCount").Int(),
-		TotalTokens:     node.Get("totalTokenCount").Int(),
-	}
-	if detail.TotalTokens == 0 {
-		detail.TotalTokens = detail.InputTokens + detail.OutputTokens + detail.ReasoningTokens
-	}
-	return detail, true
+	return parseGeminiFamilyUsageDetail(node), true
 }
 
 var stopChunkWithoutUsage sync.Map
@@ -522,12 +482,16 @@ func StripUsageMetadataFromJSON(rawJSON []byte) ([]byte, bool) {
 	cleaned := jsonBytes
 	var changed bool
 
-	if gjson.GetBytes(cleaned, "usageMetadata").Exists() {
+	if usageMetadata = gjson.GetBytes(cleaned, "usageMetadata"); usageMetadata.Exists() {
+		// Rename usageMetadata to cpaUsageMetadata in the message_start event of Claude
+		cleaned, _ = sjson.SetRawBytes(cleaned, "cpaUsageMetadata", []byte(usageMetadata.Raw))
 		cleaned, _ = sjson.DeleteBytes(cleaned, "usageMetadata")
 		changed = true
 	}
 
-	if gjson.GetBytes(cleaned, "response.usageMetadata").Exists() {
+	if usageMetadata = gjson.GetBytes(cleaned, "response.usageMetadata"); usageMetadata.Exists() {
+		// Rename usageMetadata to cpaUsageMetadata in the message_start event of Claude
+		cleaned, _ = sjson.SetRawBytes(cleaned, "response.cpaUsageMetadata", []byte(usageMetadata.Raw))
 		cleaned, _ = sjson.DeleteBytes(cleaned, "response.usageMetadata")
 		changed = true
 	}

internal/translator/antigravity/claude/antigravity_claude_request.go

@@ -86,6 +86,10 @@ func ConvertClaudeRequestToAntigravity(modelName string, inputRawJSON []byte, _
 				hasSystemInstruction = true
 			}
 		}
+	} else if systemResult.Type == gjson.String {
+		systemInstructionJSON = `{"role":"user","parts":[{"text":""}]}`
+		systemInstructionJSON, _ = sjson.Set(systemInstructionJSON, "parts.0.text", systemResult.String())
+		hasSystemInstruction = true
 	}
 
 	// contents
@@ -121,32 +125,31 @@ func ConvertClaudeRequestToAntigravity(modelName string, inputRawJSON []byte, _
 						// Use GetThinkingText to handle wrapped thinking objects
 						thinkingText := util.GetThinkingText(contentResult)
 						signatureResult := contentResult.Get("signature")
-					clientSignature := ""
-					if signatureResult.Exists() && signatureResult.String() != "" {
-						clientSignature = signatureResult.String()
-					}
-
-					// Always try cached signature first (more reliable than client-provided)
-					// Client may send stale or invalid signatures from different sessions
-					signature := ""
-					if sessionID != "" && thinkingText != "" {
-						if cachedSig := cache.GetCachedSignature(sessionID, thinkingText); cachedSig != "" {
-							signature = cachedSig
-							log.Debugf("Using cached signature for thinking block")
+						clientSignature := ""
+						if signatureResult.Exists() && signatureResult.String() != "" {
+							clientSignature = signatureResult.String()
 						}
-					}
 
-					// Fallback to client signature only if cache miss and client signature is valid
-					if signature == "" && cache.HasValidSignature(clientSignature) {
-						signature = clientSignature
-						log.Debugf("Using client-provided signature for thinking block")
-					}
+						// Always try cached signature first (more reliable than client-provided)
+						// Client may send stale or invalid signatures from different sessions
+						signature := ""
+						if sessionID != "" && thinkingText != "" {
+							if cachedSig := cache.GetCachedSignature(sessionID, thinkingText); cachedSig != "" {
+								signature = cachedSig
+								log.Debugf("Using cached signature for thinking block")
+							}
+						}
 
-					// Store for subsequent tool_use in the same message
-					if cache.HasValidSignature(signature) {
-						currentMessageThinkingSignature = signature
-					}
+						// Fallback to client signature only if cache miss and client signature is valid
+						if signature == "" && cache.HasValidSignature(clientSignature) {
+							signature = clientSignature
+							log.Debugf("Using client-provided signature for thinking block")
+						}
 
+						// Store for subsequent tool_use in the same message
+						if cache.HasValidSignature(signature) {
+							currentMessageThinkingSignature = signature
+						}
 
 						// Skip trailing unsigned thinking blocks on last assistant message
 						isUnsigned := !cache.HasValidSignature(signature)
@@ -321,6 +324,7 @@ func ConvertClaudeRequestToAntigravity(modelName string, inputRawJSON []byte, _
 	// tools
 	toolsJSON := ""
 	toolDeclCount := 0
+	allowedToolKeys := []string{"name", "description", "behavior", "parameters", "parametersJsonSchema", "response", "responseJsonSchema"}
 	toolsResult := gjson.GetBytes(rawJSON, "tools")
 	if toolsResult.IsArray() {
 		toolsJSON = `[{"functionDeclarations":[]}]`
@@ -333,10 +337,12 @@ func ConvertClaudeRequestToAntigravity(modelName string, inputRawJSON []byte, _
 				inputSchema := util.CleanJSONSchemaForAntigravity(inputSchemaResult.Raw)
 				tool, _ := sjson.Delete(toolResult.Raw, "input_schema")
 				tool, _ = sjson.SetRaw(tool, "parametersJsonSchema", inputSchema)
-				tool, _ = sjson.Delete(tool, "strict")
-				tool, _ = sjson.Delete(tool, "input_examples")
-				tool, _ = sjson.Delete(tool, "type")
-				tool, _ = sjson.Delete(tool, "cache_control")
+				for toolKey := range gjson.Parse(tool).Map() {
+					if util.InArray(allowedToolKeys, toolKey) {
+						continue
+					}
+					tool, _ = sjson.Delete(tool, toolKey)
+				}
 				toolsJSON, _ = sjson.SetRaw(toolsJSON, "0.functionDeclarations.-1", tool)
 				toolDeclCount++
 			}

internal/translator/antigravity/claude/antigravity_claude_response.go

@@ -35,6 +35,7 @@ type Params struct {
 	CandidatesTokenCount int64  // Cached candidate token count from usage metadata
 	ThoughtsTokenCount   int64  // Cached thinking token count from usage metadata
 	TotalTokenCount      int64  // Cached total token count from usage metadata
+	CachedTokenCount     int64  // Cached content token count (indicates prompt caching)
 	HasSentFinalEvents   bool   // Indicates if final content/message events have been sent
 	HasToolUse           bool   // Indicates if tool use was observed in the stream
 	HasContent           bool   // Tracks whether any content (text, thinking, or tool use) has been output
@@ -98,6 +99,14 @@ func ConvertAntigravityResponseToClaude(_ context.Context, _ string, originalReq
 		// This follows the Claude Code API specification for streaming message initialization
 		messageStartTemplate := `{"type": "message_start", "message": {"id": "msg_1nZdL29xx5MUA1yADyHTEsnR8uuvGzszyY", "type": "message", "role": "assistant", "content": [], "model": "claude-3-5-sonnet-20241022", "stop_reason": null, "stop_sequence": null, "usage": {"input_tokens": 0, "output_tokens": 0}}}`
 
+		// Use cpaUsageMetadata within the message_start event for Claude.
+		if promptTokenCount := gjson.GetBytes(rawJSON, "response.cpaUsageMetadata.promptTokenCount"); promptTokenCount.Exists() {
+			messageStartTemplate, _ = sjson.Set(messageStartTemplate, "message.usage.input_tokens", promptTokenCount.Int())
+		}
+		if candidatesTokenCount := gjson.GetBytes(rawJSON, "response.cpaUsageMetadata.candidatesTokenCount"); candidatesTokenCount.Exists() {
+			messageStartTemplate, _ = sjson.Set(messageStartTemplate, "message.usage.output_tokens", candidatesTokenCount.Int())
+		}
+
 		// Override default values with actual response metadata if available from the Gemini CLI response
 		if modelVersionResult := gjson.GetBytes(rawJSON, "response.modelVersion"); modelVersionResult.Exists() {
 			messageStartTemplate, _ = sjson.Set(messageStartTemplate, "message.model", modelVersionResult.String())
@@ -270,7 +279,8 @@ func ConvertAntigravityResponseToClaude(_ context.Context, _ string, originalReq
 
 	if usageResult := gjson.GetBytes(rawJSON, "response.usageMetadata"); usageResult.Exists() {
 		params.HasUsageMetadata = true
-		params.PromptTokenCount = usageResult.Get("promptTokenCount").Int()
+		params.CachedTokenCount = usageResult.Get("cachedContentTokenCount").Int()
+		params.PromptTokenCount = usageResult.Get("promptTokenCount").Int() - params.CachedTokenCount
 		params.CandidatesTokenCount = usageResult.Get("candidatesTokenCount").Int()
 		params.ThoughtsTokenCount = usageResult.Get("thoughtsTokenCount").Int()
 		params.TotalTokenCount = usageResult.Get("totalTokenCount").Int()
@@ -322,6 +332,14 @@ func appendFinalEvents(params *Params, output *string, force bool) {
 	*output = *output + "event: message_delta\n"
 	*output = *output + "data: "
 	delta := fmt.Sprintf(`{"type":"message_delta","delta":{"stop_reason":"%s","stop_sequence":null},"usage":{"input_tokens":%d,"output_tokens":%d}}`, stopReason, params.PromptTokenCount, usageOutputTokens)
+	// Add cache_read_input_tokens if cached tokens are present (indicates prompt caching is working)
+	if params.CachedTokenCount > 0 {
+		var err error
+		delta, err = sjson.Set(delta, "usage.cache_read_input_tokens", params.CachedTokenCount)
+		if err != nil {
+			log.Warnf("antigravity claude response: failed to set cache_read_input_tokens: %v", err)
+		}
+	}
 	*output = *output + delta + "\n\n\n"
 
 	params.HasSentFinalEvents = true
@@ -361,6 +379,7 @@ func ConvertAntigravityResponseToClaudeNonStream(_ context.Context, _ string, or
 	candidateTokens := root.Get("response.usageMetadata.candidatesTokenCount").Int()
 	thoughtTokens := root.Get("response.usageMetadata.thoughtsTokenCount").Int()
 	totalTokens := root.Get("response.usageMetadata.totalTokenCount").Int()
+	cachedTokens := root.Get("response.usageMetadata.cachedContentTokenCount").Int()
 	outputTokens := candidateTokens + thoughtTokens
 	if outputTokens == 0 && totalTokens > 0 {
 		outputTokens = totalTokens - promptTokens
@@ -374,6 +393,14 @@ func ConvertAntigravityResponseToClaudeNonStream(_ context.Context, _ string, or
 	responseJSON, _ = sjson.Set(responseJSON, "model", root.Get("response.modelVersion").String())
 	responseJSON, _ = sjson.Set(responseJSON, "usage.input_tokens", promptTokens)
 	responseJSON, _ = sjson.Set(responseJSON, "usage.output_tokens", outputTokens)
+	// Add cache_read_input_tokens if cached tokens are present (indicates prompt caching is working)
+	if cachedTokens > 0 {
+		var err error
+		responseJSON, err = sjson.Set(responseJSON, "usage.cache_read_input_tokens", cachedTokens)
+		if err != nil {
+			log.Warnf("antigravity claude response: failed to set cache_read_input_tokens: %v", err)
+		}
+	}
 
 	contentArrayInitialized := false
 	ensureContentArray := func() {
@@ -452,7 +479,7 @@ func ConvertAntigravityResponseToClaudeNonStream(_ context.Context, _ string, or
 				toolBlock, _ = sjson.Set(toolBlock, "id", fmt.Sprintf("tool_%d", toolIDCounter))
 				toolBlock, _ = sjson.Set(toolBlock, "name", name)
 
-				if args := functionCall.Get("args"); args.Exists() && args.Raw != "" && gjson.Valid(args.Raw) {
+				if args := functionCall.Get("args"); args.Exists() && args.Raw != "" && gjson.Valid(args.Raw) && args.IsObject() {
 					toolBlock, _ = sjson.SetRaw(toolBlock, "input", args.Raw)
 				}
 

internal/translator/antigravity/gemini/antigravity_gemini_request.go

@@ -7,7 +7,6 @@ package gemini
 
 import (
 	"bytes"
-	"encoding/json"
 	"fmt"
 
 	"github.com/router-for-me/CLIProxyAPI/v6/internal/translator/gemini/common"
@@ -135,41 +134,31 @@ func ConvertGeminiRequestToAntigravity(_ string, inputRawJSON []byte, _ bool) []
 
 // FunctionCallGroup represents a group of function calls and their responses
 type FunctionCallGroup struct {
-	ModelContent    map[string]interface{}
-	FunctionCalls   []gjson.Result
 	ResponsesNeeded int
 }
 
-// parseFunctionResponse attempts to unmarshal a function response part.
-// Falls back to gjson extraction if standard json.Unmarshal fails.
-func parseFunctionResponse(response gjson.Result) map[string]interface{} {
-	var responseMap map[string]interface{}
-	err := json.Unmarshal([]byte(response.Raw), &responseMap)
-	if err == nil {
-		return responseMap
+// parseFunctionResponseRaw attempts to normalize a function response part into a JSON object string.
+// Falls back to a minimal "functionResponse" object when parsing fails.
+func parseFunctionResponseRaw(response gjson.Result) string {
+	if response.IsObject() && gjson.Valid(response.Raw) {
+		return response.Raw
 	}
 
-	log.Debugf("unmarshal function response failed, using fallback: %v", err)
+	log.Debugf("parse function response failed, using fallback")
 	funcResp := response.Get("functionResponse")
 	if funcResp.Exists() {
-		fr := map[string]interface{}{
-			"name": funcResp.Get("name").String(),
-			"response": map[string]interface{}{
-				"result": funcResp.Get("response").String(),
-			},
-		}
+		fr := `{"functionResponse":{"name":"","response":{"result":""}}}`
+		fr, _ = sjson.Set(fr, "functionResponse.name", funcResp.Get("name").String())
+		fr, _ = sjson.Set(fr, "functionResponse.response.result", funcResp.Get("response").String())
 		if id := funcResp.Get("id").String(); id != "" {
-			fr["id"] = id
+			fr, _ = sjson.Set(fr, "functionResponse.id", id)
 		}
-		return map[string]interface{}{"functionResponse": fr}
+		return fr
 	}
 
-	return map[string]interface{}{
-		"functionResponse": map[string]interface{}{
-			"name":     "unknown",
-			"response": map[string]interface{}{"result": response.String()},
-		},
-	}
+	fr := `{"functionResponse":{"name":"unknown","response":{"result":""}}}`
+	fr, _ = sjson.Set(fr, "functionResponse.response.result", response.String())
+	return fr
 }
 
 // fixCLIToolResponse performs sophisticated tool response format conversion and grouping.
@@ -196,7 +185,7 @@ func fixCLIToolResponse(input string) (string, error) {
 	}
 
 	// Initialize data structures for processing and grouping
-	var newContents []interface{}          // Final processed contents array
+	contentsWrapper := `{"contents":[]}`
 	var pendingGroups []*FunctionCallGroup // Groups awaiting completion with responses
 	var collectedResponses []gjson.Result  // Standalone responses to be matched
 
@@ -228,17 +217,16 @@ func fixCLIToolResponse(input string) (string, error) {
 					collectedResponses = collectedResponses[group.ResponsesNeeded:]
 
 					// Create merged function response content
-					var responseParts []interface{}
+					functionResponseContent := `{"parts":[],"role":"function"}`
 					for _, response := range groupResponses {
-						responseParts = append(responseParts, parseFunctionResponse(response))
+						partRaw := parseFunctionResponseRaw(response)
+						if partRaw != "" {
+							functionResponseContent, _ = sjson.SetRaw(functionResponseContent, "parts.-1", partRaw)
+						}
 					}
 
-					if len(responseParts) > 0 {
-						functionResponseContent := map[string]interface{}{
-							"parts": responseParts,
-							"role":  "function",
-						}
-						newContents = append(newContents, functionResponseContent)
+					if gjson.Get(functionResponseContent, "parts.#").Int() > 0 {
+						contentsWrapper, _ = sjson.SetRaw(contentsWrapper, "contents.-1", functionResponseContent)
 					}
 
 					// Remove this group as it's been satisfied
@@ -252,50 +240,42 @@ func fixCLIToolResponse(input string) (string, error) {
 
 		// If this is a model with function calls, create a new group
 		if role == "model" {
-			var functionCallsInThisModel []gjson.Result
+			functionCallsCount := 0
 			parts.ForEach(func(_, part gjson.Result) bool {
 				if part.Get("functionCall").Exists() {
-					functionCallsInThisModel = append(functionCallsInThisModel, part)
+					functionCallsCount++
 				}
 				return true
 			})
 
-			if len(functionCallsInThisModel) > 0 {
+			if functionCallsCount > 0 {
 				// Add the model content
-				var contentMap map[string]interface{}
-				errUnmarshal := json.Unmarshal([]byte(value.Raw), &contentMap)
-				if errUnmarshal != nil {
-					log.Warnf("failed to unmarshal model content: %v\n", errUnmarshal)
+				if !value.IsObject() {
+					log.Warnf("failed to parse model content")
 					return true
 				}
-				newContents = append(newContents, contentMap)
+				contentsWrapper, _ = sjson.SetRaw(contentsWrapper, "contents.-1", value.Raw)
 
 				// Create a new group for tracking responses
 				group := &FunctionCallGroup{
-					ModelContent:    contentMap,
-					FunctionCalls:   functionCallsInThisModel,
-					ResponsesNeeded: len(functionCallsInThisModel),
+					ResponsesNeeded: functionCallsCount,
 				}
 				pendingGroups = append(pendingGroups, group)
 			} else {
 				// Regular model content without function calls
-				var contentMap map[string]interface{}
-				errUnmarshal := json.Unmarshal([]byte(value.Raw), &contentMap)
-				if errUnmarshal != nil {
-					log.Warnf("failed to unmarshal content: %v\n", errUnmarshal)
+				if !value.IsObject() {
+					log.Warnf("failed to parse content")
 					return true
 				}
-				newContents = append(newContents, contentMap)
+				contentsWrapper, _ = sjson.SetRaw(contentsWrapper, "contents.-1", value.Raw)
 			}
 		} else {
 			// Non-model content (user, etc.)
-			var contentMap map[string]interface{}
-			errUnmarshal := json.Unmarshal([]byte(value.Raw), &contentMap)
-			if errUnmarshal != nil {
-				log.Warnf("failed to unmarshal content: %v\n", errUnmarshal)
+			if !value.IsObject() {
+				log.Warnf("failed to parse content")
 				return true
 			}
-			newContents = append(newContents, contentMap)
+			contentsWrapper, _ = sjson.SetRaw(contentsWrapper, "contents.-1", value.Raw)
 		}
 
 		return true
@@ -307,25 +287,23 @@ func fixCLIToolResponse(input string) (string, error) {
 			groupResponses := collectedResponses[:group.ResponsesNeeded]
 			collectedResponses = collectedResponses[group.ResponsesNeeded:]
 
-			var responseParts []interface{}
+			functionResponseContent := `{"parts":[],"role":"function"}`
 			for _, response := range groupResponses {
-				responseParts = append(responseParts, parseFunctionResponse(response))
+				partRaw := parseFunctionResponseRaw(response)
+				if partRaw != "" {
+					functionResponseContent, _ = sjson.SetRaw(functionResponseContent, "parts.-1", partRaw)
+				}
 			}
 
-			if len(responseParts) > 0 {
-				functionResponseContent := map[string]interface{}{
-					"parts": responseParts,
-					"role":  "function",
-				}
-				newContents = append(newContents, functionResponseContent)
+			if gjson.Get(functionResponseContent, "parts.#").Int() > 0 {
+				contentsWrapper, _ = sjson.SetRaw(contentsWrapper, "contents.-1", functionResponseContent)
 			}
 		}
 	}
 
 	// Update the original JSON with the new contents
 	result := input
-	newContentsJSON, _ := json.Marshal(newContents)
-	result, _ = sjson.Set(result, "request.contents", json.RawMessage(newContentsJSON))
+	result, _ = sjson.SetRaw(result, "request.contents", gjson.Get(contentsWrapper, "contents").Raw)
 
 	return result, nil
 }

internal/translator/antigravity/openai/chat-completions/antigravity_openai_request.go

@@ -192,6 +192,14 @@ func ConvertOpenAIRequestToAntigravity(modelName string, inputRawJSON []byte, _
 				} else if content.IsObject() && content.Get("type").String() == "text" {
 					out, _ = sjson.SetBytes(out, "request.systemInstruction.role", "user")
 					out, _ = sjson.SetBytes(out, "request.systemInstruction.parts.0.text", content.Get("text").String())
+				} else if content.IsArray() {
+					contents := content.Array()
+					if len(contents) > 0 {
+						out, _ = sjson.SetBytes(out, "request.systemInstruction.role", "user")
+						for j := 0; j < len(contents); j++ {
+							out, _ = sjson.SetBytes(out, fmt.Sprintf("request.systemInstruction.parts.%d.text", j), contents[j].Get("text").String())
+						}
+					}
 				}
 			} else if role == "user" || (role == "system" && len(arr) == 1) {
 				// Build single user content node to avoid splitting into multiple contents
@@ -239,10 +247,30 @@ func ConvertOpenAIRequestToAntigravity(modelName string, inputRawJSON []byte, _
 			} else if role == "assistant" {
 				node := []byte(`{"role":"model","parts":[]}`)
 				p := 0
-				if content.Type == gjson.String {
+				if content.Type == gjson.String && content.String() != "" {
 					node, _ = sjson.SetBytes(node, "parts.-1.text", content.String())
-					out, _ = sjson.SetRawBytes(out, "request.contents.-1", node)
 					p++
+				} else if content.IsArray() {
+					// Assistant multimodal content (e.g. text + image) -> single model content with parts
+					for _, item := range content.Array() {
+						switch item.Get("type").String() {
+						case "text":
+							p++
+						case "image_url":
+							// If the assistant returned an inline data URL, preserve it for history fidelity.
+							imageURL := item.Get("image_url.url").String()
+							if len(imageURL) > 5 { // expect data:...
+								pieces := strings.SplitN(imageURL[5:], ";", 2)
+								if len(pieces) == 2 && len(pieces[1]) > 7 {
+									mime := pieces[0]
+									data := pieces[1][7:]
+									node, _ = sjson.SetBytes(node, "parts."+itoa(p)+".inlineData.mime_type", mime)
+									node, _ = sjson.SetBytes(node, "parts."+itoa(p)+".inlineData.data", data)
+									p++
+								}
+							}
+						}
+					}
 				}
 
 				// Tool calls -> single model content with functionCall parts
@@ -258,7 +286,11 @@ func ConvertOpenAIRequestToAntigravity(modelName string, inputRawJSON []byte, _
 						fargs := tc.Get("function.arguments").String()
 						node, _ = sjson.SetBytes(node, "parts."+itoa(p)+".functionCall.id", fid)
 						node, _ = sjson.SetBytes(node, "parts."+itoa(p)+".functionCall.name", fname)
-						node, _ = sjson.SetRawBytes(node, "parts."+itoa(p)+".functionCall.args", []byte(fargs))
+						if gjson.Valid(fargs) {
+							node, _ = sjson.SetRawBytes(node, "parts."+itoa(p)+".functionCall.args", []byte(fargs))
+						} else {
+							node, _ = sjson.SetBytes(node, "parts."+itoa(p)+".functionCall.args.params", []byte(fargs))
+						}
 						node, _ = sjson.SetBytes(node, "parts."+itoa(p)+".thoughtSignature", geminiCLIFunctionThoughtSignature)
 						p++
 						if fid != "" {
@@ -293,6 +325,8 @@ func ConvertOpenAIRequestToAntigravity(modelName string, inputRawJSON []byte, _
 					if pp > 0 {
 						out, _ = sjson.SetRawBytes(out, "request.contents.-1", toolNode)
 					}
+				} else {
+					out, _ = sjson.SetRawBytes(out, "request.contents.-1", node)
 				}
 			}
 		}
@@ -319,7 +353,7 @@ func ConvertOpenAIRequestToAntigravity(modelName string, inputRawJSON []byte, _
 								log.Warnf("Failed to set default schema type for tool '%s': %v", fn.Get("name").String(), errSet)
 								continue
 							}
-							fnRaw, errSet = sjson.Set(fnRaw, "parametersJsonSchema.properties", map[string]interface{}{})
+							fnRaw, errSet = sjson.SetRaw(fnRaw, "parametersJsonSchema.properties", `{}`)
 							if errSet != nil {
 								log.Warnf("Failed to set default schema properties for tool '%s': %v", fn.Get("name").String(), errSet)
 								continue
@@ -334,7 +368,7 @@ func ConvertOpenAIRequestToAntigravity(modelName string, inputRawJSON []byte, _
 							log.Warnf("Failed to set default schema type for tool '%s': %v", fn.Get("name").String(), errSet)
 							continue
 						}
-						fnRaw, errSet = sjson.Set(fnRaw, "parametersJsonSchema.properties", map[string]interface{}{})
+						fnRaw, errSet = sjson.SetRaw(fnRaw, "parametersJsonSchema.properties", `{}`)
 						if errSet != nil {
 							log.Warnf("Failed to set default schema properties for tool '%s': %v", fn.Get("name").String(), errSet)
 							continue

internal/translator/antigravity/openai/chat-completions/antigravity_openai_response.go

@@ -8,12 +8,13 @@ package chat_completions
 import (
 	"bytes"
 	"context"
-	"encoding/json"
 	"fmt"
 	"strings"
 	"sync/atomic"
 	"time"
 
+	log "github.com/sirupsen/logrus"
+
 	. "github.com/router-for-me/CLIProxyAPI/v6/internal/translator/gemini/openai/chat-completions"
 	"github.com/tidwall/gjson"
 	"github.com/tidwall/sjson"
@@ -86,18 +87,27 @@ func ConvertAntigravityResponseToOpenAI(_ context.Context, _ string, originalReq
 
 	// Extract and set usage metadata (token counts).
 	if usageResult := gjson.GetBytes(rawJSON, "response.usageMetadata"); usageResult.Exists() {
+		cachedTokenCount := usageResult.Get("cachedContentTokenCount").Int()
 		if candidatesTokenCountResult := usageResult.Get("candidatesTokenCount"); candidatesTokenCountResult.Exists() {
 			template, _ = sjson.Set(template, "usage.completion_tokens", candidatesTokenCountResult.Int())
 		}
 		if totalTokenCountResult := usageResult.Get("totalTokenCount"); totalTokenCountResult.Exists() {
 			template, _ = sjson.Set(template, "usage.total_tokens", totalTokenCountResult.Int())
 		}
-		promptTokenCount := usageResult.Get("promptTokenCount").Int()
+		promptTokenCount := usageResult.Get("promptTokenCount").Int() - cachedTokenCount
 		thoughtsTokenCount := usageResult.Get("thoughtsTokenCount").Int()
 		template, _ = sjson.Set(template, "usage.prompt_tokens", promptTokenCount+thoughtsTokenCount)
 		if thoughtsTokenCount > 0 {
 			template, _ = sjson.Set(template, "usage.completion_tokens_details.reasoning_tokens", thoughtsTokenCount)
 		}
+		// Include cached token count if present (indicates prompt caching is working)
+		if cachedTokenCount > 0 {
+			var err error
+			template, err = sjson.Set(template, "usage.prompt_tokens_details.cached_tokens", cachedTokenCount)
+			if err != nil {
+				log.Warnf("antigravity openai response: failed to set cached_tokens: %v", err)
+			}
+		}
 	}
 
 	// Process the main content part of the response.
@@ -171,21 +181,16 @@ func ConvertAntigravityResponseToOpenAI(_ context.Context, _ string, originalReq
 					mimeType = "image/png"
 				}
 				imageURL := fmt.Sprintf("data:%s;base64,%s", mimeType, data)
-				imagePayload, err := json.Marshal(map[string]any{
-					"type": "image_url",
-					"image_url": map[string]string{
-						"url": imageURL,
-					},
-				})
-				if err != nil {
-					continue
-				}
 				imagesResult := gjson.Get(template, "choices.0.delta.images")
 				if !imagesResult.Exists() || !imagesResult.IsArray() {
 					template, _ = sjson.SetRaw(template, "choices.0.delta.images", `[]`)
 				}
+				imageIndex := len(gjson.Get(template, "choices.0.delta.images").Array())
+				imagePayload := `{"type":"image_url","image_url":{"url":""}}`
+				imagePayload, _ = sjson.Set(imagePayload, "index", imageIndex)
+				imagePayload, _ = sjson.Set(imagePayload, "image_url.url", imageURL)
 				template, _ = sjson.Set(template, "choices.0.delta.role", "assistant")
-				template, _ = sjson.SetRaw(template, "choices.0.delta.images.-1", string(imagePayload))
+				template, _ = sjson.SetRaw(template, "choices.0.delta.images.-1", imagePayload)
 			}
 		}
 	}

internal/translator/claude/gemini/claude_gemini_request.go

@@ -194,7 +194,7 @@ func ConvertGeminiRequestToClaude(modelName string, inputRawJSON []byte, stream
 						if name := fc.Get("name"); name.Exists() {
 							toolUse, _ = sjson.Set(toolUse, "name", name.String())
 						}
-						if args := fc.Get("args"); args.Exists() {
+						if args := fc.Get("args"); args.Exists() && args.IsObject() {
 							toolUse, _ = sjson.SetRaw(toolUse, "input", args.Raw)
 						}
 						msg, _ = sjson.SetRaw(msg, "content.-1", toolUse)
@@ -314,11 +314,11 @@ func ConvertGeminiRequestToClaude(modelName string, inputRawJSON []byte, stream
 			if mode := funcCalling.Get("mode"); mode.Exists() {
 				switch mode.String() {
 				case "AUTO":
-					out, _ = sjson.Set(out, "tool_choice", map[string]interface{}{"type": "auto"})
+					out, _ = sjson.SetRaw(out, "tool_choice", `{"type":"auto"}`)
 				case "NONE":
-					out, _ = sjson.Set(out, "tool_choice", map[string]interface{}{"type": "none"})
+					out, _ = sjson.SetRaw(out, "tool_choice", `{"type":"none"}`)
 				case "ANY":
-					out, _ = sjson.Set(out, "tool_choice", map[string]interface{}{"type": "any"})
+					out, _ = sjson.SetRaw(out, "tool_choice", `{"type":"any"}`)
 				}
 			}
 		}

internal/translator/claude/gemini/claude_gemini_response.go

@@ -263,51 +263,6 @@ func ConvertClaudeResponseToGemini(_ context.Context, modelName string, original
 	}
 }
 
-// convertArrayToJSON converts []interface{} to JSON array string
-func convertArrayToJSON(arr []interface{}) string {
-	result := "[]"
-	for _, item := range arr {
-		switch itemData := item.(type) {
-		case map[string]interface{}:
-			itemJSON := convertMapToJSON(itemData)
-			result, _ = sjson.SetRaw(result, "-1", itemJSON)
-		case string:
-			result, _ = sjson.Set(result, "-1", itemData)
-		case bool:
-			result, _ = sjson.Set(result, "-1", itemData)
-		case float64, int, int64:
-			result, _ = sjson.Set(result, "-1", itemData)
-		default:
-			result, _ = sjson.Set(result, "-1", itemData)
-		}
-	}
-	return result
-}
-
-// convertMapToJSON converts map[string]interface{} to JSON object string
-func convertMapToJSON(m map[string]interface{}) string {
-	result := "{}"
-	for key, value := range m {
-		switch val := value.(type) {
-		case map[string]interface{}:
-			nestedJSON := convertMapToJSON(val)
-			result, _ = sjson.SetRaw(result, key, nestedJSON)
-		case []interface{}:
-			arrayJSON := convertArrayToJSON(val)
-			result, _ = sjson.SetRaw(result, key, arrayJSON)
-		case string:
-			result, _ = sjson.Set(result, key, val)
-		case bool:
-			result, _ = sjson.Set(result, key, val)
-		case float64, int, int64:
-			result, _ = sjson.Set(result, key, val)
-		default:
-			result, _ = sjson.Set(result, key, val)
-		}
-	}
-	return result
-}
-
 // ConvertClaudeResponseToGeminiNonStream converts a non-streaming Claude Code response to a non-streaming Gemini response.
 // This function processes the complete Claude Code response and transforms it into a single Gemini-compatible
 // JSON response. It handles message content, tool calls, reasoning content, and usage metadata, combining all
@@ -356,8 +311,8 @@ func ConvertClaudeResponseToGeminiNonStream(_ context.Context, modelName string,
 	}
 
 	// Process each streaming event and collect parts
-	var allParts []interface{}
-	var finalUsage map[string]interface{}
+	var allParts []string
+	var finalUsageJSON string
 	var responseID string
 	var createdAt int64
 
@@ -407,16 +362,14 @@ func ConvertClaudeResponseToGeminiNonStream(_ context.Context, modelName string,
 					if text := delta.Get("text"); text.Exists() && text.String() != "" {
 						partJSON := `{"text":""}`
 						partJSON, _ = sjson.Set(partJSON, "text", text.String())
-						part := gjson.Parse(partJSON).Value().(map[string]interface{})
-						allParts = append(allParts, part)
+						allParts = append(allParts, partJSON)
 					}
 				case "thinking_delta":
 					// Process reasoning/thinking content
 					if text := delta.Get("thinking"); text.Exists() && text.String() != "" {
 						partJSON := `{"thought":true,"text":""}`
 						partJSON, _ = sjson.Set(partJSON, "text", text.String())
-						part := gjson.Parse(partJSON).Value().(map[string]interface{})
-						allParts = append(allParts, part)
+						allParts = append(allParts, partJSON)
 					}
 				case "input_json_delta":
 					// accumulate args partial_json for this index
@@ -456,9 +409,7 @@ func ConvertClaudeResponseToGeminiNonStream(_ context.Context, modelName string,
 				if argsTrim != "" {
 					functionCallJSON, _ = sjson.SetRaw(functionCallJSON, "functionCall.args", argsTrim)
 				}
-				// Parse back to interface{} for allParts
-				functionCall := gjson.Parse(functionCallJSON).Value().(map[string]interface{})
-				allParts = append(allParts, functionCall)
+				allParts = append(allParts, functionCallJSON)
 				// cleanup used state for this index
 				if newParam.ToolUseArgs != nil {
 					delete(newParam.ToolUseArgs, idx)
@@ -501,8 +452,7 @@ func ConvertClaudeResponseToGeminiNonStream(_ context.Context, modelName string,
 				// Set traffic type (required by Gemini API)
 				usageJSON, _ = sjson.Set(usageJSON, "trafficType", "PROVISIONED_THROUGHPUT")
 
-				// Convert to map[string]interface{} using gjson
-				finalUsage = gjson.Parse(usageJSON).Value().(map[string]interface{})
+				finalUsageJSON = usageJSON
 			}
 		}
 	}
@@ -520,12 +470,16 @@ func ConvertClaudeResponseToGeminiNonStream(_ context.Context, modelName string,
 
 	// Set the consolidated parts array
 	if len(consolidatedParts) > 0 {
-		template, _ = sjson.SetRaw(template, "candidates.0.content.parts", convertToJSONString(consolidatedParts))
+		partsJSON := "[]"
+		for _, partJSON := range consolidatedParts {
+			partsJSON, _ = sjson.SetRaw(partsJSON, "-1", partJSON)
+		}
+		template, _ = sjson.SetRaw(template, "candidates.0.content.parts", partsJSON)
 	}
 
 	// Set usage metadata
-	if finalUsage != nil {
-		template, _ = sjson.SetRaw(template, "usageMetadata", convertToJSONString(finalUsage))
+	if finalUsageJSON != "" {
+		template, _ = sjson.SetRaw(template, "usageMetadata", finalUsageJSON)
 	}
 
 	return template
@@ -539,12 +493,12 @@ func GeminiTokenCount(ctx context.Context, count int64) string {
 // This function processes the parts array to combine adjacent text elements and thinking elements
 // into single consolidated parts, which results in a more readable and efficient response structure.
 // Tool calls and other non-text parts are preserved as separate elements.
-func consolidateParts(parts []interface{}) []interface{} {
+func consolidateParts(parts []string) []string {
 	if len(parts) == 0 {
 		return parts
 	}
 
-	var consolidated []interface{}
+	var consolidated []string
 	var currentTextPart strings.Builder
 	var currentThoughtPart strings.Builder
 	var hasText, hasThought bool
@@ -554,8 +508,7 @@ func consolidateParts(parts []interface{}) []interface{} {
 		if hasText && currentTextPart.Len() > 0 {
 			textPartJSON := `{"text":""}`
 			textPartJSON, _ = sjson.Set(textPartJSON, "text", currentTextPart.String())
-			textPart := gjson.Parse(textPartJSON).Value().(map[string]interface{})
-			consolidated = append(consolidated, textPart)
+			consolidated = append(consolidated, textPartJSON)
 			currentTextPart.Reset()
 			hasText = false
 		}
@@ -566,42 +519,42 @@ func consolidateParts(parts []interface{}) []interface{} {
 		if hasThought && currentThoughtPart.Len() > 0 {
 			thoughtPartJSON := `{"thought":true,"text":""}`
 			thoughtPartJSON, _ = sjson.Set(thoughtPartJSON, "text", currentThoughtPart.String())
-			thoughtPart := gjson.Parse(thoughtPartJSON).Value().(map[string]interface{})
-			consolidated = append(consolidated, thoughtPart)
+			consolidated = append(consolidated, thoughtPartJSON)
 			currentThoughtPart.Reset()
 			hasThought = false
 		}
 	}
 
-	for _, part := range parts {
-		partMap, ok := part.(map[string]interface{})
-		if !ok {
+	for _, partJSON := range parts {
+		part := gjson.Parse(partJSON)
+		if !part.Exists() || !part.IsObject() {
 			// Flush any pending parts and add this non-text part
 			flushText()
 			flushThought()
-			consolidated = append(consolidated, part)
+			consolidated = append(consolidated, partJSON)
 			continue
 		}
 
-		if thought, isThought := partMap["thought"]; isThought && thought == true {
+		thought := part.Get("thought")
+		if thought.Exists() && thought.Type == gjson.True {
 			// This is a thinking part - flush any pending text first
 			flushText() // Flush any pending text first
 
-			if text, hasTextContent := partMap["text"].(string); hasTextContent {
-				currentThoughtPart.WriteString(text)
+			if text := part.Get("text"); text.Exists() && text.Type == gjson.String {
+				currentThoughtPart.WriteString(text.String())
 				hasThought = true
 			}
-		} else if text, hasTextContent := partMap["text"].(string); hasTextContent {
+		} else if text := part.Get("text"); text.Exists() && text.Type == gjson.String {
 			// This is a regular text part - flush any pending thought first
 			flushThought() // Flush any pending thought first
 
-			currentTextPart.WriteString(text)
+			currentTextPart.WriteString(text.String())
 			hasText = true
 		} else {
 			// This is some other type of part (like function call) - flush both text and thought
 			flushText()
 			flushThought()
-			consolidated = append(consolidated, part)
+			consolidated = append(consolidated, partJSON)
 		}
 	}
 
@@ -611,20 +564,3 @@ func consolidateParts(parts []interface{}) []interface{} {
 
 	return consolidated
 }
-
-// convertToJSONString converts interface{} to JSON string using sjson/gjson.
-// This function provides a consistent way to serialize different data types to JSON strings
-// for inclusion in the Gemini API response structure.
-func convertToJSONString(v interface{}) string {
-	switch val := v.(type) {
-	case []interface{}:
-		return convertArrayToJSON(val)
-	case map[string]interface{}:
-		return convertMapToJSON(val)
-	default:
-		// For simple types, create a temporary JSON and extract the value
-		temp := `{"temp":null}`
-		temp, _ = sjson.Set(temp, "temp", val)
-		return gjson.Get(temp, "temp").Raw
-	}
-}

internal/translator/claude/openai/chat-completions/claude_openai_request.go

@@ -10,7 +10,6 @@ import (
 	"crypto/rand"
 	"crypto/sha256"
 	"encoding/hex"
-	"encoding/json"
 	"fmt"
 	"math/big"
 	"strings"
@@ -137,9 +136,6 @@ func ConvertOpenAIRequestToClaude(modelName string, inputRawJSON []byte, stream
 	out, _ = sjson.Set(out, "stream", stream)
 
 	// Process messages and transform them to Claude Code format
-	var anthropicMessages []interface{}
-	var toolCallIDs []string // Track tool call IDs for matching with tool results
-
 	if messages := root.Get("messages"); messages.Exists() && messages.IsArray() {
 		messages.ForEach(func(_, message gjson.Result) bool {
 			role := message.Get("role").String()
@@ -152,33 +148,23 @@ func ConvertOpenAIRequestToClaude(modelName string, inputRawJSON []byte, stream
 					role = "user"
 				}
 
-				msg := map[string]interface{}{
-					"role":    role,
-					"content": []interface{}{},
-				}
+				msg := `{"role":"","content":[]}`
+				msg, _ = sjson.Set(msg, "role", role)
 
 				// Handle content based on its type (string or array)
 				if contentResult.Exists() && contentResult.Type == gjson.String && contentResult.String() != "" {
-					// Simple text content conversion
-					msg["content"] = []interface{}{
-						map[string]interface{}{
-							"type": "text",
-							"text": contentResult.String(),
-						},
-					}
+					part := `{"type":"text","text":""}`
+					part, _ = sjson.Set(part, "text", contentResult.String())
+					msg, _ = sjson.SetRaw(msg, "content.-1", part)
 				} else if contentResult.Exists() && contentResult.IsArray() {
-					// Array of content parts processing
-					var contentParts []interface{}
 					contentResult.ForEach(func(_, part gjson.Result) bool {
 						partType := part.Get("type").String()
 
 						switch partType {
 						case "text":
-							// Text part conversion
-							contentParts = append(contentParts, map[string]interface{}{
-								"type": "text",
-								"text": part.Get("text").String(),
-							})
+							textPart := `{"type":"text","text":""}`
+							textPart, _ = sjson.Set(textPart, "text", part.Get("text").String())
+							msg, _ = sjson.SetRaw(msg, "content.-1", textPart)
 
 						case "image_url":
 							// Convert OpenAI image format to Claude Code format
@@ -191,132 +177,95 @@ func ConvertOpenAIRequestToClaude(modelName string, inputRawJSON []byte, stream
 									mediaType := strings.TrimPrefix(mediaTypePart, "data:")
 									data := parts[1]
 
-									contentParts = append(contentParts, map[string]interface{}{
-										"type": "image",
-										"source": map[string]interface{}{
-											"type":       "base64",
-											"media_type": mediaType,
-											"data":       data,
-										},
-									})
+									imagePart := `{"type":"image","source":{"type":"base64","media_type":"","data":""}}`
+									imagePart, _ = sjson.Set(imagePart, "source.media_type", mediaType)
+									imagePart, _ = sjson.Set(imagePart, "source.data", data)
+									msg, _ = sjson.SetRaw(msg, "content.-1", imagePart)
 								}
 							}
 						}
 						return true
 					})
-					if len(contentParts) > 0 {
-						msg["content"] = contentParts
-					}
-				} else {
-					// Initialize empty content array for tool calls
-					msg["content"] = []interface{}{}
 				}
 
 				// Handle tool calls (for assistant messages)
 				if toolCalls := message.Get("tool_calls"); toolCalls.Exists() && toolCalls.IsArray() && role == "assistant" {
-					var contentParts []interface{}
-
-					// Add existing text content if any
-					if existingContent, ok := msg["content"].([]interface{}); ok {
-						contentParts = existingContent
-					}
-
 					toolCalls.ForEach(func(_, toolCall gjson.Result) bool {
 						if toolCall.Get("type").String() == "function" {
 							toolCallID := toolCall.Get("id").String()
 							if toolCallID == "" {
 								toolCallID = genToolCallID()
 							}
-							toolCallIDs = append(toolCallIDs, toolCallID)
 
 							function := toolCall.Get("function")
-							toolUse := map[string]interface{}{
-								"type": "tool_use",
-								"id":   toolCallID,
-								"name": function.Get("name").String(),
-							}
+							toolUse := `{"type":"tool_use","id":"","name":"","input":{}}`
+							toolUse, _ = sjson.Set(toolUse, "id", toolCallID)
+							toolUse, _ = sjson.Set(toolUse, "name", function.Get("name").String())
 
 							// Parse arguments for the tool call
 							if args := function.Get("arguments"); args.Exists() {
 								argsStr := args.String()
-								if argsStr != "" {
-									var argsMap map[string]interface{}
-									if err := json.Unmarshal([]byte(argsStr), &argsMap); err == nil {
-										toolUse["input"] = argsMap
+								if argsStr != "" && gjson.Valid(argsStr) {
+									argsJSON := gjson.Parse(argsStr)
+									if argsJSON.IsObject() {
+										toolUse, _ = sjson.SetRaw(toolUse, "input", argsJSON.Raw)
 									} else {
-										toolUse["input"] = map[string]interface{}{}
+										toolUse, _ = sjson.SetRaw(toolUse, "input", "{}")
 									}
 								} else {
-									toolUse["input"] = map[string]interface{}{}
+									toolUse, _ = sjson.SetRaw(toolUse, "input", "{}")
 								}
 							} else {
-								toolUse["input"] = map[string]interface{}{}
+								toolUse, _ = sjson.SetRaw(toolUse, "input", "{}")
 							}
 
-							contentParts = append(contentParts, toolUse)
+							msg, _ = sjson.SetRaw(msg, "content.-1", toolUse)
 						}
 						return true
 					})
-					msg["content"] = contentParts
 				}
 
-				anthropicMessages = append(anthropicMessages, msg)
+				out, _ = sjson.SetRaw(out, "messages.-1", msg)
 
 			case "tool":
 				// Handle tool result messages conversion
 				toolCallID := message.Get("tool_call_id").String()
 				content := message.Get("content").String()
 
-				// Create tool result message in Claude Code format
-				msg := map[string]interface{}{
-					"role": "user",
-					"content": []interface{}{
-						map[string]interface{}{
-							"type":        "tool_result",
-							"tool_use_id": toolCallID,
-							"content":     content,
-						},
-					},
-				}
-
-				anthropicMessages = append(anthropicMessages, msg)
+				msg := `{"role":"user","content":[{"type":"tool_result","tool_use_id":"","content":""}]}`
+				msg, _ = sjson.Set(msg, "content.0.tool_use_id", toolCallID)
+				msg, _ = sjson.Set(msg, "content.0.content", content)
+				out, _ = sjson.SetRaw(out, "messages.-1", msg)
 			}
 			return true
 		})
 	}
 
-	// Set messages in the output template
-	if len(anthropicMessages) > 0 {
-		messagesJSON, _ := json.Marshal(anthropicMessages)
-		out, _ = sjson.SetRaw(out, "messages", string(messagesJSON))
-	}
-
 	// Tools mapping: OpenAI tools -> Claude Code tools
 	if tools := root.Get("tools"); tools.Exists() && tools.IsArray() && len(tools.Array()) > 0 {
-		var anthropicTools []interface{}
+		hasAnthropicTools := false
 		tools.ForEach(func(_, tool gjson.Result) bool {
 			if tool.Get("type").String() == "function" {
 				function := tool.Get("function")
-				anthropicTool := map[string]interface{}{
-					"name":        function.Get("name").String(),
-					"description": function.Get("description").String(),
-				}
+				anthropicTool := `{"name":"","description":""}`
+				anthropicTool, _ = sjson.Set(anthropicTool, "name", function.Get("name").String())
+				anthropicTool, _ = sjson.Set(anthropicTool, "description", function.Get("description").String())
 
 				// Convert parameters schema for the tool
 				if parameters := function.Get("parameters"); parameters.Exists() {
-					anthropicTool["input_schema"] = parameters.Value()
-				} else if parameters = function.Get("parametersJsonSchema"); parameters.Exists() {
-					anthropicTool["input_schema"] = parameters.Value()
+					anthropicTool, _ = sjson.SetRaw(anthropicTool, "input_schema", parameters.Raw)
+				} else if parameters := function.Get("parametersJsonSchema"); parameters.Exists() {
+					anthropicTool, _ = sjson.SetRaw(anthropicTool, "input_schema", parameters.Raw)
 				}
 
-				anthropicTools = append(anthropicTools, anthropicTool)
+				out, _ = sjson.SetRaw(out, "tools.-1", anthropicTool)
+				hasAnthropicTools = true
 			}
 			return true
 		})
 
-		if len(anthropicTools) > 0 {
-			toolsJSON, _ := json.Marshal(anthropicTools)
-			out, _ = sjson.SetRaw(out, "tools", string(toolsJSON))
+		if !hasAnthropicTools {
+			out, _ = sjson.Delete(out, "tools")
 		}
 	}
 
@@ -329,18 +278,17 @@ func ConvertOpenAIRequestToClaude(modelName string, inputRawJSON []byte, stream
 			case "none":
 				// Don't set tool_choice, Claude Code will not use tools
 			case "auto":
-				out, _ = sjson.Set(out, "tool_choice", map[string]interface{}{"type": "auto"})
+				out, _ = sjson.SetRaw(out, "tool_choice", `{"type":"auto"}`)
 			case "required":
-				out, _ = sjson.Set(out, "tool_choice", map[string]interface{}{"type": "any"})
+				out, _ = sjson.SetRaw(out, "tool_choice", `{"type":"any"}`)
 			}
 		case gjson.JSON:
 			// Specific tool choice mapping
 			if toolChoice.Get("type").String() == "function" {
 				functionName := toolChoice.Get("function.name").String()
-				out, _ = sjson.Set(out, "tool_choice", map[string]interface{}{
-					"type": "tool",
-					"name": functionName,
-				})
+				toolChoiceJSON := `{"type":"tool","name":""}`
+				toolChoiceJSON, _ = sjson.Set(toolChoiceJSON, "name", functionName)
+				out, _ = sjson.SetRaw(out, "tool_choice", toolChoiceJSON)
 			}
 		default:
 		}

internal/translator/claude/openai/chat-completions/claude_openai_response.go

@@ -8,7 +8,7 @@ package chat_completions
 import (
 	"bytes"
 	"context"
-	"encoding/json"
+	"fmt"
 	"strings"
 	"time"
 
@@ -182,18 +182,11 @@ func ConvertClaudeResponseToOpenAI(_ context.Context, modelName string, original
 				if arguments == "" {
 					arguments = "{}"
 				}
-
-				toolCall := map[string]interface{}{
-					"index": index,
-					"id":    accumulator.ID,
-					"type":  "function",
-					"function": map[string]interface{}{
-						"name":      accumulator.Name,
-						"arguments": arguments,
-					},
-				}
-
-				template, _ = sjson.Set(template, "choices.0.delta.tool_calls", []interface{}{toolCall})
+				template, _ = sjson.Set(template, "choices.0.delta.tool_calls.0.index", index)
+				template, _ = sjson.Set(template, "choices.0.delta.tool_calls.0.id", accumulator.ID)
+				template, _ = sjson.Set(template, "choices.0.delta.tool_calls.0.type", "function")
+				template, _ = sjson.Set(template, "choices.0.delta.tool_calls.0.function.name", accumulator.Name)
+				template, _ = sjson.Set(template, "choices.0.delta.tool_calls.0.function.arguments", arguments)
 
 				// Clean up the accumulator for this index
 				delete((*param).(*ConvertAnthropicResponseToOpenAIParams).ToolCallsAccumulator, index)
@@ -214,12 +207,14 @@ func ConvertClaudeResponseToOpenAI(_ context.Context, modelName string, original
 
 		// Handle usage information for token counts
 		if usage := root.Get("usage"); usage.Exists() {
-			usageObj := map[string]interface{}{
-				"prompt_tokens":     usage.Get("input_tokens").Int(),
-				"completion_tokens": usage.Get("output_tokens").Int(),
-				"total_tokens":      usage.Get("input_tokens").Int() + usage.Get("output_tokens").Int(),
-			}
-			template, _ = sjson.Set(template, "usage", usageObj)
+			inputTokens := usage.Get("input_tokens").Int()
+			outputTokens := usage.Get("output_tokens").Int()
+			cacheReadInputTokens := usage.Get("cache_read_input_tokens").Int()
+			cacheCreationInputTokens := usage.Get("cache_creation_input_tokens").Int()
+			template, _ = sjson.Set(template, "usage.prompt_tokens", inputTokens+cacheCreationInputTokens)
+			template, _ = sjson.Set(template, "usage.completion_tokens", outputTokens)
+			template, _ = sjson.Set(template, "usage.total_tokens", inputTokens+outputTokens)
+			template, _ = sjson.Set(template, "usage.prompt_tokens_details.cached_tokens", cacheReadInputTokens)
 		}
 		return []string{template}
 
@@ -234,14 +229,10 @@ func ConvertClaudeResponseToOpenAI(_ context.Context, modelName string, original
 	case "error":
 		// Error event - format and return error response
 		if errorData := root.Get("error"); errorData.Exists() {
-			errorResponse := map[string]interface{}{
-				"error": map[string]interface{}{
-					"message": errorData.Get("message").String(),
-					"type":    errorData.Get("type").String(),
-				},
-			}
-			errorJSON, _ := json.Marshal(errorResponse)
-			return []string{string(errorJSON)}
+			errorJSON := `{"error":{"message":"","type":""}}`
+			errorJSON, _ = sjson.Set(errorJSON, "error.message", errorData.Get("message").String())
+			errorJSON, _ = sjson.Set(errorJSON, "error.type", errorData.Get("type").String())
+			return []string{errorJSON}
 		}
 		return []string{}
 
@@ -297,15 +288,10 @@ func ConvertClaudeResponseToOpenAINonStream(_ context.Context, _ string, origina
 	var messageID string
 	var model string
 	var createdAt int64
-	var inputTokens, outputTokens int64
-	var reasoningTokens int64
 	var stopReason string
 	var contentParts []string
 	var reasoningParts []string
-	// Use map to track tool calls by index for proper merging
-	toolCallsMap := make(map[int]map[string]interface{})
-	// Track tool call arguments accumulation
-	toolCallArgsMap := make(map[int]strings.Builder)
+	toolCallsAccumulator := make(map[int]*ToolCallAccumulator)
 
 	for _, chunk := range chunks {
 		root := gjson.ParseBytes(chunk)
@@ -318,9 +304,6 @@ func ConvertClaudeResponseToOpenAINonStream(_ context.Context, _ string, origina
 				messageID = message.Get("id").String()
 				model = message.Get("model").String()
 				createdAt = time.Now().Unix()
-				if usage := message.Get("usage"); usage.Exists() {
-					inputTokens = usage.Get("input_tokens").Int()
-				}
 			}
 
 		case "content_block_start":
@@ -331,18 +314,12 @@ func ConvertClaudeResponseToOpenAINonStream(_ context.Context, _ string, origina
 					// Start of thinking/reasoning content - skip for now as it's handled in delta
 					continue
 				} else if blockType == "tool_use" {
-					// Initialize tool call tracking for this index
+					// Initialize tool call accumulator for this index
 					index := int(root.Get("index").Int())
-					toolCallsMap[index] = map[string]interface{}{
-						"id":   contentBlock.Get("id").String(),
-						"type": "function",
-						"function": map[string]interface{}{
-							"name":      contentBlock.Get("name").String(),
-							"arguments": "",
-						},
+					toolCallsAccumulator[index] = &ToolCallAccumulator{
+						ID:   contentBlock.Get("id").String(),
+						Name: contentBlock.Get("name").String(),
 					}
-					// Initialize arguments builder for this tool call
-					toolCallArgsMap[index] = strings.Builder{}
 				}
 			}
 
@@ -365,9 +342,8 @@ func ConvertClaudeResponseToOpenAINonStream(_ context.Context, _ string, origina
 					// Accumulate tool call arguments
 					if partialJSON := delta.Get("partial_json"); partialJSON.Exists() {
 						index := int(root.Get("index").Int())
-						if builder, exists := toolCallArgsMap[index]; exists {
-							builder.WriteString(partialJSON.String())
-							toolCallArgsMap[index] = builder
+						if accumulator, exists := toolCallsAccumulator[index]; exists {
+							accumulator.Arguments.WriteString(partialJSON.String())
 						}
 					}
 				}
@@ -376,14 +352,9 @@ func ConvertClaudeResponseToOpenAINonStream(_ context.Context, _ string, origina
 		case "content_block_stop":
 			// Finalize tool call arguments for this index when content block ends
 			index := int(root.Get("index").Int())
-			if toolCall, exists := toolCallsMap[index]; exists {
-				if builder, argsExists := toolCallArgsMap[index]; argsExists {
-					// Set the accumulated arguments for the tool call
-					arguments := builder.String()
-					if arguments == "" {
-						arguments = "{}"
-					}
-					toolCall["function"].(map[string]interface{})["arguments"] = arguments
+			if accumulator, exists := toolCallsAccumulator[index]; exists {
+				if accumulator.Arguments.Len() == 0 {
+					accumulator.Arguments.WriteString("{}")
 				}
 			}
 
@@ -395,11 +366,14 @@ func ConvertClaudeResponseToOpenAINonStream(_ context.Context, _ string, origina
 				}
 			}
 			if usage := root.Get("usage"); usage.Exists() {
-				outputTokens = usage.Get("output_tokens").Int()
-				// Estimate reasoning tokens from accumulated thinking content
-				if len(reasoningParts) > 0 {
-					reasoningTokens = int64(len(strings.Join(reasoningParts, "")) / 4) // Rough estimation
-				}
+				inputTokens := usage.Get("input_tokens").Int()
+				outputTokens := usage.Get("output_tokens").Int()
+				cacheReadInputTokens := usage.Get("cache_read_input_tokens").Int()
+				cacheCreationInputTokens := usage.Get("cache_creation_input_tokens").Int()
+				out, _ = sjson.Set(out, "usage.prompt_tokens", inputTokens+cacheCreationInputTokens)
+				out, _ = sjson.Set(out, "usage.completion_tokens", outputTokens)
+				out, _ = sjson.Set(out, "usage.total_tokens", inputTokens+outputTokens)
+				out, _ = sjson.Set(out, "usage.prompt_tokens_details.cached_tokens", cacheReadInputTokens)
 			}
 		}
 	}
@@ -421,24 +395,35 @@ func ConvertClaudeResponseToOpenAINonStream(_ context.Context, _ string, origina
 	}
 
 	// Set tool calls if any were accumulated during processing
-	if len(toolCallsMap) > 0 {
-		// Convert tool calls map to array, preserving order by index
-		var toolCallsArray []interface{}
-		// Find the maximum index to determine the range
+	if len(toolCallsAccumulator) > 0 {
+		toolCallsCount := 0
 		maxIndex := -1
-		for index := range toolCallsMap {
+		for index := range toolCallsAccumulator {
 			if index > maxIndex {
 				maxIndex = index
 			}
 		}
-		// Iterate through all possible indices up to maxIndex
+
 		for i := 0; i <= maxIndex; i++ {
-			if toolCall, exists := toolCallsMap[i]; exists {
-				toolCallsArray = append(toolCallsArray, toolCall)
+			accumulator, exists := toolCallsAccumulator[i]
+			if !exists {
+				continue
 			}
+
+			arguments := accumulator.Arguments.String()
+
+			idPath := fmt.Sprintf("choices.0.message.tool_calls.%d.id", toolCallsCount)
+			typePath := fmt.Sprintf("choices.0.message.tool_calls.%d.type", toolCallsCount)
+			namePath := fmt.Sprintf("choices.0.message.tool_calls.%d.function.name", toolCallsCount)
+			argumentsPath := fmt.Sprintf("choices.0.message.tool_calls.%d.function.arguments", toolCallsCount)
+
+			out, _ = sjson.Set(out, idPath, accumulator.ID)
+			out, _ = sjson.Set(out, typePath, "function")
+			out, _ = sjson.Set(out, namePath, accumulator.Name)
+			out, _ = sjson.Set(out, argumentsPath, arguments)
+			toolCallsCount++
 		}
-		if len(toolCallsArray) > 0 {
-			out, _ = sjson.Set(out, "choices.0.message.tool_calls", toolCallsArray)
+		if toolCallsCount > 0 {
 			out, _ = sjson.Set(out, "choices.0.finish_reason", "tool_calls")
 		} else {
 			out, _ = sjson.Set(out, "choices.0.finish_reason", mapAnthropicStopReasonToOpenAI(stopReason))
@@ -447,16 +432,5 @@ func ConvertClaudeResponseToOpenAINonStream(_ context.Context, _ string, origina
 		out, _ = sjson.Set(out, "choices.0.finish_reason", mapAnthropicStopReasonToOpenAI(stopReason))
 	}
 
-	// Set usage information including prompt tokens, completion tokens, and total tokens
-	totalTokens := inputTokens + outputTokens
-	out, _ = sjson.Set(out, "usage.prompt_tokens", inputTokens)
-	out, _ = sjson.Set(out, "usage.completion_tokens", outputTokens)
-	out, _ = sjson.Set(out, "usage.total_tokens", totalTokens)
-
-	// Add reasoning tokens to usage details if any reasoning content was processed
-	if reasoningTokens > 0 {
-		out, _ = sjson.Set(out, "usage.completion_tokens_details.reasoning_tokens", reasoningTokens)
-	}
-
 	return out
 }

internal/translator/claude/openai/responses/claude_openai-responses_request.go

@@ -114,13 +114,16 @@ func ConvertOpenAIResponsesRequestToClaude(modelName string, inputRawJSON []byte
 					var builder strings.Builder
 					if parts := item.Get("content"); parts.Exists() && parts.IsArray() {
 						parts.ForEach(func(_, part gjson.Result) bool {
-							text := part.Get("text").String()
+							textResult := part.Get("text")
+							text := textResult.String()
 							if builder.Len() > 0 && text != "" {
 								builder.WriteByte('\n')
 							}
 							builder.WriteString(text)
 							return true
 						})
+					} else if parts.Type == gjson.String {
+						builder.WriteString(parts.String())
 					}
 					instructionsText = builder.String()
 					if instructionsText != "" {
@@ -207,6 +210,8 @@ func ConvertOpenAIResponsesRequestToClaude(modelName string, inputRawJSON []byte
 						}
 						return true
 					})
+				} else if parts.Type == gjson.String {
+					textAggregate.WriteString(parts.String())
 				}
 
 				// Fallback to given role if content types not decisive
@@ -254,7 +259,10 @@ func ConvertOpenAIResponsesRequestToClaude(modelName string, inputRawJSON []byte
 				toolUse, _ = sjson.Set(toolUse, "id", callID)
 				toolUse, _ = sjson.Set(toolUse, "name", name)
 				if argsStr != "" && gjson.Valid(argsStr) {
-					toolUse, _ = sjson.SetRaw(toolUse, "input", argsStr)
+					argsJSON := gjson.Parse(argsStr)
+					if argsJSON.IsObject() {
+						toolUse, _ = sjson.SetRaw(toolUse, "input", argsJSON.Raw)
+					}
 				}
 
 				asst := `{"role":"assistant","content":[]}`
@@ -309,16 +317,18 @@ func ConvertOpenAIResponsesRequestToClaude(modelName string, inputRawJSON []byte
 		case gjson.String:
 			switch toolChoice.String() {
 			case "auto":
-				out, _ = sjson.Set(out, "tool_choice", map[string]interface{}{"type": "auto"})
+				out, _ = sjson.SetRaw(out, "tool_choice", `{"type":"auto"}`)
 			case "none":
 				// Leave unset; implies no tools
 			case "required":
-				out, _ = sjson.Set(out, "tool_choice", map[string]interface{}{"type": "any"})
+				out, _ = sjson.SetRaw(out, "tool_choice", `{"type":"any"}`)
 			}
 		case gjson.JSON:
 			if toolChoice.Get("type").String() == "function" {
 				fn := toolChoice.Get("function.name").String()
-				out, _ = sjson.Set(out, "tool_choice", map[string]interface{}{"type": "tool", "name": fn})
+				toolChoiceJSON := `{"name":"","type":"tool"}`
+				toolChoiceJSON, _ = sjson.Set(toolChoiceJSON, "name", fn)
+				out, _ = sjson.SetRaw(out, "tool_choice", toolChoiceJSON)
 			}
 		default:
 

internal/translator/claude/openai/responses/claude_openai-responses_response.go

@@ -344,31 +344,20 @@ func ConvertClaudeResponseToOpenAIResponses(ctx context.Context, modelName strin
 		}
 
 		// Build response.output from aggregated state
-		var outputs []interface{}
+		outputsWrapper := `{"arr":[]}`
 		// reasoning item (if any)
 		if st.ReasoningBuf.Len() > 0 || st.ReasoningPartAdded {
-			r := map[string]interface{}{
-				"id":      st.ReasoningItemID,
-				"type":    "reasoning",
-				"summary": []interface{}{map[string]interface{}{"type": "summary_text", "text": st.ReasoningBuf.String()}},
-			}
-			outputs = append(outputs, r)
+			item := `{"id":"","type":"reasoning","summary":[{"type":"summary_text","text":""}]}`
+			item, _ = sjson.Set(item, "id", st.ReasoningItemID)
+			item, _ = sjson.Set(item, "summary.0.text", st.ReasoningBuf.String())
+			outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 		}
 		// assistant message item (if any text)
 		if st.TextBuf.Len() > 0 || st.InTextBlock || st.CurrentMsgID != "" {
-			m := map[string]interface{}{
-				"id":     st.CurrentMsgID,
-				"type":   "message",
-				"status": "completed",
-				"content": []interface{}{map[string]interface{}{
-					"type":        "output_text",
-					"annotations": []interface{}{},
-					"logprobs":    []interface{}{},
-					"text":        st.TextBuf.String(),
-				}},
-				"role": "assistant",
-			}
-			outputs = append(outputs, m)
+			item := `{"id":"","type":"message","status":"completed","content":[{"type":"output_text","annotations":[],"logprobs":[],"text":""}],"role":"assistant"}`
+			item, _ = sjson.Set(item, "id", st.CurrentMsgID)
+			item, _ = sjson.Set(item, "content.0.text", st.TextBuf.String())
+			outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 		}
 		// function_call items (in ascending index order for determinism)
 		if len(st.FuncArgsBuf) > 0 {
@@ -395,19 +384,16 @@ func ConvertClaudeResponseToOpenAIResponses(ctx context.Context, modelName strin
 				if callID == "" && st.CurrentFCID != "" {
 					callID = st.CurrentFCID
 				}
-				item := map[string]interface{}{
-					"id":        fmt.Sprintf("fc_%s", callID),
-					"type":      "function_call",
-					"status":    "completed",
-					"arguments": args,
-					"call_id":   callID,
-					"name":      name,
-				}
-				outputs = append(outputs, item)
+				item := `{"id":"","type":"function_call","status":"completed","arguments":"","call_id":"","name":""}`
+				item, _ = sjson.Set(item, "id", fmt.Sprintf("fc_%s", callID))
+				item, _ = sjson.Set(item, "arguments", args)
+				item, _ = sjson.Set(item, "call_id", callID)
+				item, _ = sjson.Set(item, "name", name)
+				outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 			}
 		}
-		if len(outputs) > 0 {
-			completed, _ = sjson.Set(completed, "response.output", outputs)
+		if gjson.Get(outputsWrapper, "arr.#").Int() > 0 {
+			completed, _ = sjson.SetRaw(completed, "response.output", gjson.Get(outputsWrapper, "arr").Raw)
 		}
 
 		reasoningTokens := int64(0)
@@ -628,27 +614,18 @@ func ConvertClaudeResponseToOpenAIResponsesNonStream(_ context.Context, _ string
 	}
 
 	// Build output array
-	var outputs []interface{}
+	outputsWrapper := `{"arr":[]}`
 	if reasoningBuf.Len() > 0 {
-		outputs = append(outputs, map[string]interface{}{
-			"id":      reasoningItemID,
-			"type":    "reasoning",
-			"summary": []interface{}{map[string]interface{}{"type": "summary_text", "text": reasoningBuf.String()}},
-		})
+		item := `{"id":"","type":"reasoning","summary":[{"type":"summary_text","text":""}]}`
+		item, _ = sjson.Set(item, "id", reasoningItemID)
+		item, _ = sjson.Set(item, "summary.0.text", reasoningBuf.String())
+		outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 	}
 	if currentMsgID != "" || textBuf.Len() > 0 {
-		outputs = append(outputs, map[string]interface{}{
-			"id":     currentMsgID,
-			"type":   "message",
-			"status": "completed",
-			"content": []interface{}{map[string]interface{}{
-				"type":        "output_text",
-				"annotations": []interface{}{},
-				"logprobs":    []interface{}{},
-				"text":        textBuf.String(),
-			}},
-			"role": "assistant",
-		})
+		item := `{"id":"","type":"message","status":"completed","content":[{"type":"output_text","annotations":[],"logprobs":[],"text":""}],"role":"assistant"}`
+		item, _ = sjson.Set(item, "id", currentMsgID)
+		item, _ = sjson.Set(item, "content.0.text", textBuf.String())
+		outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 	}
 	if len(toolCalls) > 0 {
 		// Preserve index order
@@ -669,18 +646,16 @@ func ConvertClaudeResponseToOpenAIResponsesNonStream(_ context.Context, _ string
 			if args == "" {
 				args = "{}"
 			}
-			outputs = append(outputs, map[string]interface{}{
-				"id":        fmt.Sprintf("fc_%s", st.id),
-				"type":      "function_call",
-				"status":    "completed",
-				"arguments": args,
-				"call_id":   st.id,
-				"name":      st.name,
-			})
+			item := `{"id":"","type":"function_call","status":"completed","arguments":"","call_id":"","name":""}`
+			item, _ = sjson.Set(item, "id", fmt.Sprintf("fc_%s", st.id))
+			item, _ = sjson.Set(item, "arguments", args)
+			item, _ = sjson.Set(item, "call_id", st.id)
+			item, _ = sjson.Set(item, "name", st.name)
+			outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 		}
 	}
-	if len(outputs) > 0 {
-		out, _ = sjson.Set(out, "output", outputs)
+	if gjson.Get(outputsWrapper, "arr.#").Int() > 0 {
+		out, _ = sjson.SetRaw(out, "output", gjson.Get(outputsWrapper, "arr").Raw)
 	}
 
 	// Usage

internal/translator/codex/claude/codex_claude_response.go

@@ -9,7 +9,6 @@ package claude
 import (
 	"bytes"
 	"context"
-	"encoding/json"
 	"fmt"
 	"strings"
 
@@ -191,21 +190,12 @@ func ConvertCodexResponseToClaudeNonStream(_ context.Context, _ string, original
 		return ""
 	}
 
-	response := map[string]interface{}{
-		"id":            responseData.Get("id").String(),
-		"type":          "message",
-		"role":          "assistant",
-		"model":         responseData.Get("model").String(),
-		"content":       []interface{}{},
-		"stop_reason":   nil,
-		"stop_sequence": nil,
-		"usage": map[string]interface{}{
-			"input_tokens":  responseData.Get("usage.input_tokens").Int(),
-			"output_tokens": responseData.Get("usage.output_tokens").Int(),
-		},
-	}
+	out := `{"id":"","type":"message","role":"assistant","model":"","content":[],"stop_reason":null,"stop_sequence":null,"usage":{"input_tokens":0,"output_tokens":0}}`
+	out, _ = sjson.Set(out, "id", responseData.Get("id").String())
+	out, _ = sjson.Set(out, "model", responseData.Get("model").String())
+	out, _ = sjson.Set(out, "usage.input_tokens", responseData.Get("usage.input_tokens").Int())
+	out, _ = sjson.Set(out, "usage.output_tokens", responseData.Get("usage.output_tokens").Int())
 
-	var contentBlocks []interface{}
 	hasToolCall := false
 
 	if output := responseData.Get("output"); output.Exists() && output.IsArray() {
@@ -244,10 +234,9 @@ func ConvertCodexResponseToClaudeNonStream(_ context.Context, _ string, original
 					}
 				}
 				if thinkingBuilder.Len() > 0 {
-					contentBlocks = append(contentBlocks, map[string]interface{}{
-						"type":     "thinking",
-						"thinking": thinkingBuilder.String(),
-					})
+					block := `{"type":"thinking","thinking":""}`
+					block, _ = sjson.Set(block, "thinking", thinkingBuilder.String())
+					out, _ = sjson.SetRaw(out, "content.-1", block)
 				}
 			case "message":
 				if content := item.Get("content"); content.Exists() {
@@ -256,10 +245,9 @@ func ConvertCodexResponseToClaudeNonStream(_ context.Context, _ string, original
 							if part.Get("type").String() == "output_text" {
 								text := part.Get("text").String()
 								if text != "" {
-									contentBlocks = append(contentBlocks, map[string]interface{}{
-										"type": "text",
-										"text": text,
-									})
+									block := `{"type":"text","text":""}`
+									block, _ = sjson.Set(block, "text", text)
+									out, _ = sjson.SetRaw(out, "content.-1", block)
 								}
 							}
 							return true
@@ -267,10 +255,9 @@ func ConvertCodexResponseToClaudeNonStream(_ context.Context, _ string, original
 					} else {
 						text := content.String()
 						if text != "" {
-							contentBlocks = append(contentBlocks, map[string]interface{}{
-								"type": "text",
-								"text": text,
-							})
+							block := `{"type":"text","text":""}`
+							block, _ = sjson.Set(block, "text", text)
+							out, _ = sjson.SetRaw(out, "content.-1", block)
 						}
 					}
 				}
@@ -281,54 +268,41 @@ func ConvertCodexResponseToClaudeNonStream(_ context.Context, _ string, original
 					name = original
 				}
 
-				toolBlock := map[string]interface{}{
-					"type":  "tool_use",
-					"id":    item.Get("call_id").String(),
-					"name":  name,
-					"input": map[string]interface{}{},
-				}
-
-				if argsStr := item.Get("arguments").String(); argsStr != "" {
-					var args interface{}
-					if err := json.Unmarshal([]byte(argsStr), &args); err == nil {
-						toolBlock["input"] = args
+				toolBlock := `{"type":"tool_use","id":"","name":"","input":{}}`
+				toolBlock, _ = sjson.Set(toolBlock, "id", item.Get("call_id").String())
+				toolBlock, _ = sjson.Set(toolBlock, "name", name)
+				inputRaw := "{}"
+				if argsStr := item.Get("arguments").String(); argsStr != "" && gjson.Valid(argsStr) {
+					argsJSON := gjson.Parse(argsStr)
+					if argsJSON.IsObject() {
+						inputRaw = argsJSON.Raw
 					}
 				}
-
-				contentBlocks = append(contentBlocks, toolBlock)
+				toolBlock, _ = sjson.SetRaw(toolBlock, "input", inputRaw)
+				out, _ = sjson.SetRaw(out, "content.-1", toolBlock)
 			}
 			return true
 		})
 	}
 
-	if len(contentBlocks) > 0 {
-		response["content"] = contentBlocks
-	}
-
 	if stopReason := responseData.Get("stop_reason"); stopReason.Exists() && stopReason.String() != "" {
-		response["stop_reason"] = stopReason.String()
+		out, _ = sjson.Set(out, "stop_reason", stopReason.String())
 	} else if hasToolCall {
-		response["stop_reason"] = "tool_use"
+		out, _ = sjson.Set(out, "stop_reason", "tool_use")
 	} else {
-		response["stop_reason"] = "end_turn"
+		out, _ = sjson.Set(out, "stop_reason", "end_turn")
 	}
 
 	if stopSequence := responseData.Get("stop_sequence"); stopSequence.Exists() && stopSequence.String() != "" {
-		response["stop_sequence"] = stopSequence.Value()
+		out, _ = sjson.SetRaw(out, "stop_sequence", stopSequence.Raw)
 	}
 
 	if responseData.Get("usage.input_tokens").Exists() || responseData.Get("usage.output_tokens").Exists() {
-		response["usage"] = map[string]interface{}{
-			"input_tokens":  responseData.Get("usage.input_tokens").Int(),
-			"output_tokens": responseData.Get("usage.output_tokens").Int(),
-		}
+		out, _ = sjson.Set(out, "usage.input_tokens", responseData.Get("usage.input_tokens").Int())
+		out, _ = sjson.Set(out, "usage.output_tokens", responseData.Get("usage.output_tokens").Int())
 	}
 
-	responseJSON, err := json.Marshal(response)
-	if err != nil {
-		return ""
-	}
-	return string(responseJSON)
+	return out
 }
 
 // buildReverseMapFromClaudeOriginalShortToOriginal builds a map[short]original from original Claude request tools.

internal/translator/codex/gemini/codex_gemini_response.go

@@ -7,7 +7,6 @@ package gemini
 import (
 	"bytes"
 	"context"
-	"encoding/json"
 	"fmt"
 	"time"
 
@@ -190,19 +189,19 @@ func ConvertCodexResponseToGeminiNonStream(_ context.Context, modelName string,
 		}
 
 		// Process output content to build parts array
-		var parts []interface{}
 		hasToolCall := false
-		var pendingFunctionCalls []interface{}
+		var pendingFunctionCalls []string
 
 		flushPendingFunctionCalls := func() {
-			if len(pendingFunctionCalls) > 0 {
-				// Add all pending function calls as individual parts
-				// This maintains the original Gemini API format while ensuring consecutive calls are grouped together
-				for _, fc := range pendingFunctionCalls {
-					parts = append(parts, fc)
-				}
-				pendingFunctionCalls = nil
+			if len(pendingFunctionCalls) == 0 {
+				return
 			}
+			// Add all pending function calls as individual parts
+			// This maintains the original Gemini API format while ensuring consecutive calls are grouped together
+			for _, fc := range pendingFunctionCalls {
+				template, _ = sjson.SetRaw(template, "candidates.0.content.parts.-1", fc)
+			}
+			pendingFunctionCalls = nil
 		}
 
 		if output := responseData.Get("output"); output.Exists() && output.IsArray() {
@@ -216,11 +215,9 @@ func ConvertCodexResponseToGeminiNonStream(_ context.Context, modelName string,
 
 					// Add thinking content
 					if content := value.Get("content"); content.Exists() {
-						part := map[string]interface{}{
-							"thought": true,
-							"text":    content.String(),
-						}
-						parts = append(parts, part)
+						part := `{"text":"","thought":true}`
+						part, _ = sjson.Set(part, "text", content.String())
+						template, _ = sjson.SetRaw(template, "candidates.0.content.parts.-1", part)
 					}
 
 				case "message":
@@ -232,10 +229,9 @@ func ConvertCodexResponseToGeminiNonStream(_ context.Context, modelName string,
 						content.ForEach(func(_, contentItem gjson.Result) bool {
 							if contentItem.Get("type").String() == "output_text" {
 								if text := contentItem.Get("text"); text.Exists() {
-									part := map[string]interface{}{
-										"text": text.String(),
-									}
-									parts = append(parts, part)
+									part := `{"text":""}`
+									part, _ = sjson.Set(part, "text", text.String())
+									template, _ = sjson.SetRaw(template, "candidates.0.content.parts.-1", part)
 								}
 							}
 							return true
@@ -245,28 +241,21 @@ func ConvertCodexResponseToGeminiNonStream(_ context.Context, modelName string,
 				case "function_call":
 					// Collect function call for potential merging with consecutive ones
 					hasToolCall = true
-					functionCall := map[string]interface{}{
-						"functionCall": map[string]interface{}{
-							"name": func() string {
-								n := value.Get("name").String()
-								rev := buildReverseMapFromGeminiOriginal(originalRequestRawJSON)
-								if orig, ok := rev[n]; ok {
-									return orig
-								}
-								return n
-							}(),
-							"args": map[string]interface{}{},
-						},
+					functionCall := `{"functionCall":{"args":{},"name":""}}`
+					{
+						n := value.Get("name").String()
+						rev := buildReverseMapFromGeminiOriginal(originalRequestRawJSON)
+						if orig, ok := rev[n]; ok {
+							n = orig
+						}
+						functionCall, _ = sjson.Set(functionCall, "functionCall.name", n)
 					}
 
 					// Parse and set arguments
 					if argsStr := value.Get("arguments").String(); argsStr != "" {
 						argsResult := gjson.Parse(argsStr)
 						if argsResult.IsObject() {
-							var args map[string]interface{}
-							if err := json.Unmarshal([]byte(argsStr), &args); err == nil {
-								functionCall["functionCall"].(map[string]interface{})["args"] = args
-							}
+							functionCall, _ = sjson.SetRaw(functionCall, "functionCall.args", argsStr)
 						}
 					}
 
@@ -279,11 +268,6 @@ func ConvertCodexResponseToGeminiNonStream(_ context.Context, modelName string,
 			flushPendingFunctionCalls()
 		}
 
-		// Set the parts array
-		if len(parts) > 0 {
-			template, _ = sjson.SetRaw(template, "candidates.0.content.parts", mustMarshalJSON(parts))
-		}
-
 		// Set finish reason based on whether there were tool calls
 		if hasToolCall {
 			template, _ = sjson.Set(template, "candidates.0.finishReason", "STOP")
@@ -323,15 +307,6 @@ func buildReverseMapFromGeminiOriginal(original []byte) map[string]string {
 	return rev
 }
 
-// mustMarshalJSON marshals a value to JSON, panicking on error.
-func mustMarshalJSON(v interface{}) string {
-	data, err := json.Marshal(v)
-	if err != nil {
-		return ""
-	}
-	return string(data)
-}
-
 func GeminiTokenCount(ctx context.Context, count int64) string {
 	return fmt.Sprintf(`{"totalTokens":%d,"promptTokensDetails":[{"modality":"TEXT","tokenCount":%d}]}`, count, count)
 }

internal/translator/gemini-cli/claude/gemini-cli_claude_request.go

@@ -7,10 +7,8 @@ package claude
 
 import (
 	"bytes"
-	"encoding/json"
 	"strings"
 
-	client "github.com/router-for-me/CLIProxyAPI/v6/internal/interfaces"
 	"github.com/router-for-me/CLIProxyAPI/v6/internal/translator/gemini/common"
 	"github.com/router-for-me/CLIProxyAPI/v6/internal/util"
 	"github.com/tidwall/gjson"
@@ -41,92 +39,102 @@ func ConvertClaudeRequestToCLI(modelName string, inputRawJSON []byte, _ bool) []
 	rawJSON := bytes.Clone(inputRawJSON)
 	rawJSON = bytes.Replace(rawJSON, []byte(`"url":{"type":"string","format":"uri",`), []byte(`"url":{"type":"string",`), -1)
 
+	// Build output Gemini CLI request JSON
+	out := `{"model":"","request":{"contents":[]}}`
+	out, _ = sjson.Set(out, "model", modelName)
+
 	// system instruction
-	var systemInstruction *client.Content
-	systemResult := gjson.GetBytes(rawJSON, "system")
-	if systemResult.IsArray() {
-		systemResults := systemResult.Array()
-		systemInstruction = &client.Content{Role: "user", Parts: []client.Part{}}
-		for i := 0; i < len(systemResults); i++ {
-			systemPromptResult := systemResults[i]
-			systemTypePromptResult := systemPromptResult.Get("type")
-			if systemTypePromptResult.Type == gjson.String && systemTypePromptResult.String() == "text" {
-				systemPrompt := systemPromptResult.Get("text").String()
-				systemPart := client.Part{Text: systemPrompt}
-				systemInstruction.Parts = append(systemInstruction.Parts, systemPart)
+	if systemResult := gjson.GetBytes(rawJSON, "system"); systemResult.IsArray() {
+		systemInstruction := `{"role":"user","parts":[]}`
+		hasSystemParts := false
+		systemResult.ForEach(func(_, systemPromptResult gjson.Result) bool {
+			if systemPromptResult.Get("type").String() == "text" {
+				textResult := systemPromptResult.Get("text")
+				if textResult.Type == gjson.String {
+					part := `{"text":""}`
+					part, _ = sjson.Set(part, "text", textResult.String())
+					systemInstruction, _ = sjson.SetRaw(systemInstruction, "parts.-1", part)
+					hasSystemParts = true
+				}
 			}
+			return true
+		})
+		if hasSystemParts {
+			out, _ = sjson.SetRaw(out, "request.systemInstruction", systemInstruction)
 		}
-		if len(systemInstruction.Parts) == 0 {
-			systemInstruction = nil
-		}
+	} else if systemResult.Type == gjson.String {
+		out, _ = sjson.Set(out, "request.systemInstruction.parts.-1.text", systemResult.String())
 	}
 
 	// contents
-	contents := make([]client.Content, 0)
-	messagesResult := gjson.GetBytes(rawJSON, "messages")
-	if messagesResult.IsArray() {
-		messageResults := messagesResult.Array()
-		for i := 0; i < len(messageResults); i++ {
-			messageResult := messageResults[i]
+	if messagesResult := gjson.GetBytes(rawJSON, "messages"); messagesResult.IsArray() {
+		messagesResult.ForEach(func(_, messageResult gjson.Result) bool {
 			roleResult := messageResult.Get("role")
 			if roleResult.Type != gjson.String {
-				continue
+				return true
 			}
 			role := roleResult.String()
 			if role == "assistant" {
 				role = "model"
 			}
-			clientContent := client.Content{Role: role, Parts: []client.Part{}}
+
+			contentJSON := `{"role":"","parts":[]}`
+			contentJSON, _ = sjson.Set(contentJSON, "role", role)
+
 			contentsResult := messageResult.Get("content")
 			if contentsResult.IsArray() {
-				contentResults := contentsResult.Array()
-				for j := 0; j < len(contentResults); j++ {
-					contentResult := contentResults[j]
-					contentTypeResult := contentResult.Get("type")
-					if contentTypeResult.Type == gjson.String && contentTypeResult.String() == "text" {
-						prompt := contentResult.Get("text").String()
-						clientContent.Parts = append(clientContent.Parts, client.Part{Text: prompt})
-					} else if contentTypeResult.Type == gjson.String && contentTypeResult.String() == "tool_use" {
+				contentsResult.ForEach(func(_, contentResult gjson.Result) bool {
+					switch contentResult.Get("type").String() {
+					case "text":
+						part := `{"text":""}`
+						part, _ = sjson.Set(part, "text", contentResult.Get("text").String())
+						contentJSON, _ = sjson.SetRaw(contentJSON, "parts.-1", part)
+
+					case "tool_use":
 						functionName := contentResult.Get("name").String()
 						functionArgs := contentResult.Get("input").String()
-						var args map[string]any
-						if err := json.Unmarshal([]byte(functionArgs), &args); err == nil {
-							clientContent.Parts = append(clientContent.Parts, client.Part{
-								FunctionCall:     &client.FunctionCall{Name: functionName, Args: args},
-								ThoughtSignature: geminiCLIClaudeThoughtSignature,
-							})
+						argsResult := gjson.Parse(functionArgs)
+						if argsResult.IsObject() && gjson.Valid(functionArgs) {
+							part := `{"thoughtSignature":"","functionCall":{"name":"","args":{}}}`
+							part, _ = sjson.Set(part, "thoughtSignature", geminiCLIClaudeThoughtSignature)
+							part, _ = sjson.Set(part, "functionCall.name", functionName)
+							part, _ = sjson.SetRaw(part, "functionCall.args", functionArgs)
+							contentJSON, _ = sjson.SetRaw(contentJSON, "parts.-1", part)
 						}
-					} else if contentTypeResult.Type == gjson.String && contentTypeResult.String() == "tool_result" {
+
+					case "tool_result":
 						toolCallID := contentResult.Get("tool_use_id").String()
-						if toolCallID != "" {
-							funcName := toolCallID
-							toolCallIDs := strings.Split(toolCallID, "-")
-							if len(toolCallIDs) > 1 {
-								funcName = strings.Join(toolCallIDs[0:len(toolCallIDs)-1], "-")
-							}
-							responseData := contentResult.Get("content").Raw
-							functionResponse := client.FunctionResponse{Name: funcName, Response: map[string]interface{}{"result": responseData}}
-							clientContent.Parts = append(clientContent.Parts, client.Part{FunctionResponse: &functionResponse})
+						if toolCallID == "" {
+							return true
 						}
+						funcName := toolCallID
+						toolCallIDs := strings.Split(toolCallID, "-")
+						if len(toolCallIDs) > 1 {
+							funcName = strings.Join(toolCallIDs[0:len(toolCallIDs)-1], "-")
+						}
+						responseData := contentResult.Get("content").Raw
+						part := `{"functionResponse":{"name":"","response":{"result":""}}}`
+						part, _ = sjson.Set(part, "functionResponse.name", funcName)
+						part, _ = sjson.Set(part, "functionResponse.response.result", responseData)
+						contentJSON, _ = sjson.SetRaw(contentJSON, "parts.-1", part)
 					}
-				}
-				contents = append(contents, clientContent)
+					return true
+				})
+				out, _ = sjson.SetRaw(out, "request.contents.-1", contentJSON)
 			} else if contentsResult.Type == gjson.String {
-				prompt := contentsResult.String()
-				contents = append(contents, client.Content{Role: role, Parts: []client.Part{{Text: prompt}}})
+				part := `{"text":""}`
+				part, _ = sjson.Set(part, "text", contentsResult.String())
+				contentJSON, _ = sjson.SetRaw(contentJSON, "parts.-1", part)
+				out, _ = sjson.SetRaw(out, "request.contents.-1", contentJSON)
 			}
-		}
+			return true
+		})
 	}
 
 	// tools
-	var tools []client.ToolDeclaration
-	toolsResult := gjson.GetBytes(rawJSON, "tools")
-	if toolsResult.IsArray() {
-		tools = make([]client.ToolDeclaration, 1)
-		tools[0].FunctionDeclarations = make([]any, 0)
-		toolsResults := toolsResult.Array()
-		for i := 0; i < len(toolsResults); i++ {
-			toolResult := toolsResults[i]
+	if toolsResult := gjson.GetBytes(rawJSON, "tools"); toolsResult.IsArray() {
+		hasTools := false
+		toolsResult.ForEach(func(_, toolResult gjson.Result) bool {
 			inputSchemaResult := toolResult.Get("input_schema")
 			if inputSchemaResult.Exists() && inputSchemaResult.IsObject() {
 				inputSchema := inputSchemaResult.Raw
@@ -136,30 +144,19 @@ func ConvertClaudeRequestToCLI(modelName string, inputRawJSON []byte, _ bool) []
 				tool, _ = sjson.Delete(tool, "input_examples")
 				tool, _ = sjson.Delete(tool, "type")
 				tool, _ = sjson.Delete(tool, "cache_control")
-				var toolDeclaration any
-				if err := json.Unmarshal([]byte(tool), &toolDeclaration); err == nil {
-					tools[0].FunctionDeclarations = append(tools[0].FunctionDeclarations, toolDeclaration)
+				if gjson.Valid(tool) && gjson.Parse(tool).IsObject() {
+					if !hasTools {
+						out, _ = sjson.SetRaw(out, "request.tools", `[{"functionDeclarations":[]}]`)
+						hasTools = true
+					}
+					out, _ = sjson.SetRaw(out, "request.tools.0.functionDeclarations.-1", tool)
 				}
 			}
+			return true
+		})
+		if !hasTools {
+			out, _ = sjson.Delete(out, "request.tools")
 		}
-	} else {
-		tools = make([]client.ToolDeclaration, 0)
-	}
-
-	// Build output Gemini CLI request JSON
-	out := `{"model":"","request":{"contents":[]}}`
-	out, _ = sjson.Set(out, "model", modelName)
-	if systemInstruction != nil {
-		b, _ := json.Marshal(systemInstruction)
-		out, _ = sjson.SetRaw(out, "request.systemInstruction", string(b))
-	}
-	if len(contents) > 0 {
-		b, _ := json.Marshal(contents)
-		out, _ = sjson.SetRaw(out, "request.contents", string(b))
-	}
-	if len(tools) > 0 && len(tools[0].FunctionDeclarations) > 0 {
-		b, _ := json.Marshal(tools)
-		out, _ = sjson.SetRaw(out, "request.tools", string(b))
 	}
 
 	// Map Anthropic thinking -> Gemini thinkingBudget/include_thoughts when type==enabled

internal/translator/gemini-cli/claude/gemini-cli_claude_response.go

@@ -9,7 +9,6 @@ package claude
 import (
 	"bytes"
 	"context"
-	"encoding/json"
 	"fmt"
 	"strings"
 	"sync/atomic"
@@ -276,22 +275,16 @@ func ConvertGeminiCLIResponseToClaudeNonStream(_ context.Context, _ string, orig
 
 	root := gjson.ParseBytes(rawJSON)
 
-	response := map[string]interface{}{
-		"id":            root.Get("response.responseId").String(),
-		"type":          "message",
-		"role":          "assistant",
-		"model":         root.Get("response.modelVersion").String(),
-		"content":       []interface{}{},
-		"stop_reason":   nil,
-		"stop_sequence": nil,
-		"usage": map[string]interface{}{
-			"input_tokens":  root.Get("response.usageMetadata.promptTokenCount").Int(),
-			"output_tokens": root.Get("response.usageMetadata.candidatesTokenCount").Int() + root.Get("response.usageMetadata.thoughtsTokenCount").Int(),
-		},
-	}
+	out := `{"id":"","type":"message","role":"assistant","model":"","content":[],"stop_reason":null,"stop_sequence":null,"usage":{"input_tokens":0,"output_tokens":0}}`
+	out, _ = sjson.Set(out, "id", root.Get("response.responseId").String())
+	out, _ = sjson.Set(out, "model", root.Get("response.modelVersion").String())
+
+	inputTokens := root.Get("response.usageMetadata.promptTokenCount").Int()
+	outputTokens := root.Get("response.usageMetadata.candidatesTokenCount").Int() + root.Get("response.usageMetadata.thoughtsTokenCount").Int()
+	out, _ = sjson.Set(out, "usage.input_tokens", inputTokens)
+	out, _ = sjson.Set(out, "usage.output_tokens", outputTokens)
 
 	parts := root.Get("response.candidates.0.content.parts")
-	var contentBlocks []interface{}
 	textBuilder := strings.Builder{}
 	thinkingBuilder := strings.Builder{}
 	toolIDCounter := 0
@@ -301,10 +294,9 @@ func ConvertGeminiCLIResponseToClaudeNonStream(_ context.Context, _ string, orig
 		if textBuilder.Len() == 0 {
 			return
 		}
-		contentBlocks = append(contentBlocks, map[string]interface{}{
-			"type": "text",
-			"text": textBuilder.String(),
-		})
+		block := `{"type":"text","text":""}`
+		block, _ = sjson.Set(block, "text", textBuilder.String())
+		out, _ = sjson.SetRaw(out, "content.-1", block)
 		textBuilder.Reset()
 	}
 
@@ -312,10 +304,9 @@ func ConvertGeminiCLIResponseToClaudeNonStream(_ context.Context, _ string, orig
 		if thinkingBuilder.Len() == 0 {
 			return
 		}
-		contentBlocks = append(contentBlocks, map[string]interface{}{
-			"type":     "thinking",
-			"thinking": thinkingBuilder.String(),
-		})
+		block := `{"type":"thinking","thinking":""}`
+		block, _ = sjson.Set(block, "thinking", thinkingBuilder.String())
+		out, _ = sjson.SetRaw(out, "content.-1", block)
 		thinkingBuilder.Reset()
 	}
 
@@ -339,21 +330,15 @@ func ConvertGeminiCLIResponseToClaudeNonStream(_ context.Context, _ string, orig
 
 				name := functionCall.Get("name").String()
 				toolIDCounter++
-				toolBlock := map[string]interface{}{
-					"type":  "tool_use",
-					"id":    fmt.Sprintf("tool_%d", toolIDCounter),
-					"name":  name,
-					"input": map[string]interface{}{},
+				toolBlock := `{"type":"tool_use","id":"","name":"","input":{}}`
+				toolBlock, _ = sjson.Set(toolBlock, "id", fmt.Sprintf("tool_%d", toolIDCounter))
+				toolBlock, _ = sjson.Set(toolBlock, "name", name)
+				inputRaw := "{}"
+				if args := functionCall.Get("args"); args.Exists() && gjson.Valid(args.Raw) && args.IsObject() {
+					inputRaw = args.Raw
 				}
-
-				if args := functionCall.Get("args"); args.Exists() {
-					var parsed interface{}
-					if err := json.Unmarshal([]byte(args.Raw), &parsed); err == nil {
-						toolBlock["input"] = parsed
-					}
-				}
-
-				contentBlocks = append(contentBlocks, toolBlock)
+				toolBlock, _ = sjson.SetRaw(toolBlock, "input", inputRaw)
+				out, _ = sjson.SetRaw(out, "content.-1", toolBlock)
 				continue
 			}
 		}
@@ -362,8 +347,6 @@ func ConvertGeminiCLIResponseToClaudeNonStream(_ context.Context, _ string, orig
 	flushThinking()
 	flushText()
 
-	response["content"] = contentBlocks
-
 	stopReason := "end_turn"
 	if hasToolCall {
 		stopReason = "tool_use"
@@ -379,19 +362,13 @@ func ConvertGeminiCLIResponseToClaudeNonStream(_ context.Context, _ string, orig
 			}
 		}
 	}
-	response["stop_reason"] = stopReason
+	out, _ = sjson.Set(out, "stop_reason", stopReason)
 
-	if usage := response["usage"].(map[string]interface{}); usage["input_tokens"] == int64(0) && usage["output_tokens"] == int64(0) {
-		if usageMeta := root.Get("response.usageMetadata"); !usageMeta.Exists() {
-			delete(response, "usage")
-		}
+	if inputTokens == int64(0) && outputTokens == int64(0) && !root.Get("response.usageMetadata").Exists() {
+		out, _ = sjson.Delete(out, "usage")
 	}
 
-	encoded, err := json.Marshal(response)
-	if err != nil {
-		return ""
-	}
-	return string(encoded)
+	return out
 }
 
 func ClaudeTokenCount(ctx context.Context, count int64) string {

internal/translator/gemini-cli/gemini/gemini-cli_gemini_request.go

@@ -7,7 +7,6 @@ package gemini
 
 import (
 	"bytes"
-	"encoding/json"
 	"fmt"
 
 	"github.com/router-for-me/CLIProxyAPI/v6/internal/translator/gemini/common"
@@ -117,8 +116,6 @@ func ConvertGeminiRequestToGeminiCLI(_ string, inputRawJSON []byte, _ bool) []by
 
 // FunctionCallGroup represents a group of function calls and their responses
 type FunctionCallGroup struct {
-	ModelContent    map[string]interface{}
-	FunctionCalls   []gjson.Result
 	ResponsesNeeded int
 }
 
@@ -146,7 +143,7 @@ func fixCLIToolResponse(input string) (string, error) {
 	}
 
 	// Initialize data structures for processing and grouping
-	var newContents []interface{}          // Final processed contents array
+	contentsWrapper := `{"contents":[]}`
 	var pendingGroups []*FunctionCallGroup // Groups awaiting completion with responses
 	var collectedResponses []gjson.Result  // Standalone responses to be matched
 
@@ -178,23 +175,17 @@ func fixCLIToolResponse(input string) (string, error) {
 					collectedResponses = collectedResponses[group.ResponsesNeeded:]
 
 					// Create merged function response content
-					var responseParts []interface{}
+					functionResponseContent := `{"parts":[],"role":"function"}`
 					for _, response := range groupResponses {
-						var responseMap map[string]interface{}
-						errUnmarshal := json.Unmarshal([]byte(response.Raw), &responseMap)
-						if errUnmarshal != nil {
-							log.Warnf("failed to unmarshal function response: %v\n", errUnmarshal)
+						if !response.IsObject() {
+							log.Warnf("failed to parse function response")
 							continue
 						}
-						responseParts = append(responseParts, responseMap)
+						functionResponseContent, _ = sjson.SetRaw(functionResponseContent, "parts.-1", response.Raw)
 					}
 
-					if len(responseParts) > 0 {
-						functionResponseContent := map[string]interface{}{
-							"parts": responseParts,
-							"role":  "function",
-						}
-						newContents = append(newContents, functionResponseContent)
+					if gjson.Get(functionResponseContent, "parts.#").Int() > 0 {
+						contentsWrapper, _ = sjson.SetRaw(contentsWrapper, "contents.-1", functionResponseContent)
 					}
 
 					// Remove this group as it's been satisfied
@@ -208,50 +199,42 @@ func fixCLIToolResponse(input string) (string, error) {
 
 		// If this is a model with function calls, create a new group
 		if role == "model" {
-			var functionCallsInThisModel []gjson.Result
+			functionCallsCount := 0
 			parts.ForEach(func(_, part gjson.Result) bool {
 				if part.Get("functionCall").Exists() {
-					functionCallsInThisModel = append(functionCallsInThisModel, part)
+					functionCallsCount++
 				}
 				return true
 			})
 
-			if len(functionCallsInThisModel) > 0 {
+			if functionCallsCount > 0 {
 				// Add the model content
-				var contentMap map[string]interface{}
-				errUnmarshal := json.Unmarshal([]byte(value.Raw), &contentMap)
-				if errUnmarshal != nil {
-					log.Warnf("failed to unmarshal model content: %v\n", errUnmarshal)
+				if !value.IsObject() {
+					log.Warnf("failed to parse model content")
 					return true
 				}
-				newContents = append(newContents, contentMap)
+				contentsWrapper, _ = sjson.SetRaw(contentsWrapper, "contents.-1", value.Raw)
 
 				// Create a new group for tracking responses
 				group := &FunctionCallGroup{
-					ModelContent:    contentMap,
-					FunctionCalls:   functionCallsInThisModel,
-					ResponsesNeeded: len(functionCallsInThisModel),
+					ResponsesNeeded: functionCallsCount,
 				}
 				pendingGroups = append(pendingGroups, group)
 			} else {
 				// Regular model content without function calls
-				var contentMap map[string]interface{}
-				errUnmarshal := json.Unmarshal([]byte(value.Raw), &contentMap)
-				if errUnmarshal != nil {
-					log.Warnf("failed to unmarshal content: %v\n", errUnmarshal)
+				if !value.IsObject() {
+					log.Warnf("failed to parse content")
 					return true
 				}
-				newContents = append(newContents, contentMap)
+				contentsWrapper, _ = sjson.SetRaw(contentsWrapper, "contents.-1", value.Raw)
 			}
 		} else {
 			// Non-model content (user, etc.)
-			var contentMap map[string]interface{}
-			errUnmarshal := json.Unmarshal([]byte(value.Raw), &contentMap)
-			if errUnmarshal != nil {
-				log.Warnf("failed to unmarshal content: %v\n", errUnmarshal)
+			if !value.IsObject() {
+				log.Warnf("failed to parse content")
 				return true
 			}
-			newContents = append(newContents, contentMap)
+			contentsWrapper, _ = sjson.SetRaw(contentsWrapper, "contents.-1", value.Raw)
 		}
 
 		return true
@@ -263,31 +246,24 @@ func fixCLIToolResponse(input string) (string, error) {
 			groupResponses := collectedResponses[:group.ResponsesNeeded]
 			collectedResponses = collectedResponses[group.ResponsesNeeded:]
 
-			var responseParts []interface{}
+			functionResponseContent := `{"parts":[],"role":"function"}`
 			for _, response := range groupResponses {
-				var responseMap map[string]interface{}
-				errUnmarshal := json.Unmarshal([]byte(response.Raw), &responseMap)
-				if errUnmarshal != nil {
-					log.Warnf("failed to unmarshal function response: %v\n", errUnmarshal)
+				if !response.IsObject() {
+					log.Warnf("failed to parse function response")
 					continue
 				}
-				responseParts = append(responseParts, responseMap)
+				functionResponseContent, _ = sjson.SetRaw(functionResponseContent, "parts.-1", response.Raw)
 			}
 
-			if len(responseParts) > 0 {
-				functionResponseContent := map[string]interface{}{
-					"parts": responseParts,
-					"role":  "function",
-				}
-				newContents = append(newContents, functionResponseContent)
+			if gjson.Get(functionResponseContent, "parts.#").Int() > 0 {
+				contentsWrapper, _ = sjson.SetRaw(contentsWrapper, "contents.-1", functionResponseContent)
 			}
 		}
 	}
 
 	// Update the original JSON with the new contents
 	result := input
-	newContentsJSON, _ := json.Marshal(newContents)
-	result, _ = sjson.Set(result, "request.contents", json.RawMessage(newContentsJSON))
+	result, _ = sjson.SetRaw(result, "request.contents", gjson.Get(contentsWrapper, "contents").Raw)
 
 	return result, nil
 }

internal/translator/gemini-cli/openai/chat-completions/gemini-cli_openai_request.go

@@ -160,6 +160,14 @@ func ConvertOpenAIRequestToGeminiCLI(modelName string, inputRawJSON []byte, _ bo
 				} else if content.IsObject() && content.Get("type").String() == "text" {
 					out, _ = sjson.SetBytes(out, "request.systemInstruction.role", "user")
 					out, _ = sjson.SetBytes(out, "request.systemInstruction.parts.0.text", content.Get("text").String())
+				} else if content.IsArray() {
+					contents := content.Array()
+					if len(contents) > 0 {
+						out, _ = sjson.SetBytes(out, "request.systemInstruction.role", "user")
+						for j := 0; j < len(contents); j++ {
+							out, _ = sjson.SetBytes(out, fmt.Sprintf("request.systemInstruction.parts.%d.text", j), contents[j].Get("text").String())
+						}
+					}
 				}
 			} else if role == "user" || (role == "system" && len(arr) == 1) {
 				// Build single user content node to avoid splitting into multiple contents
@@ -210,8 +218,29 @@ func ConvertOpenAIRequestToGeminiCLI(modelName string, inputRawJSON []byte, _ bo
 				if content.Type == gjson.String {
 					// Assistant text -> single model content
 					node, _ = sjson.SetBytes(node, "parts.-1.text", content.String())
-					out, _ = sjson.SetRawBytes(out, "request.contents.-1", node)
 					p++
+				} else if content.IsArray() {
+					// Assistant multimodal content (e.g. text + image) -> single model content with parts
+					for _, item := range content.Array() {
+						switch item.Get("type").String() {
+						case "text":
+							node, _ = sjson.SetBytes(node, "parts."+itoa(p)+".text", item.Get("text").String())
+							p++
+						case "image_url":
+							// If the assistant returned an inline data URL, preserve it for history fidelity.
+							imageURL := item.Get("image_url.url").String()
+							if len(imageURL) > 5 { // expect data:...
+								pieces := strings.SplitN(imageURL[5:], ";", 2)
+								if len(pieces) == 2 && len(pieces[1]) > 7 {
+									mime := pieces[0]
+									data := pieces[1][7:]
+									node, _ = sjson.SetBytes(node, "parts."+itoa(p)+".inlineData.mime_type", mime)
+									node, _ = sjson.SetBytes(node, "parts."+itoa(p)+".inlineData.data", data)
+									p++
+								}
+							}
+						}
+					}
 				}
 
 				// Tool calls -> single model content with functionCall parts
@@ -236,7 +265,7 @@ func ConvertOpenAIRequestToGeminiCLI(modelName string, inputRawJSON []byte, _ bo
 					out, _ = sjson.SetRawBytes(out, "request.contents.-1", node)
 
 					// Append a single tool content combining name + response per function
-					toolNode := []byte(`{"role":"tool","parts":[]}`)
+					toolNode := []byte(`{"role":"user","parts":[]}`)
 					pp := 0
 					for _, fid := range fIDs {
 						if name, ok := tcID2Name[fid]; ok {
@@ -252,6 +281,8 @@ func ConvertOpenAIRequestToGeminiCLI(modelName string, inputRawJSON []byte, _ bo
 					if pp > 0 {
 						out, _ = sjson.SetRawBytes(out, "request.contents.-1", toolNode)
 					}
+				} else {
+					out, _ = sjson.SetRawBytes(out, "request.contents.-1", node)
 				}
 			}
 		}
@@ -278,7 +309,7 @@ func ConvertOpenAIRequestToGeminiCLI(modelName string, inputRawJSON []byte, _ bo
 								log.Warnf("Failed to set default schema type for tool '%s': %v", fn.Get("name").String(), errSet)
 								continue
 							}
-							fnRaw, errSet = sjson.Set(fnRaw, "parametersJsonSchema.properties", map[string]interface{}{})
+							fnRaw, errSet = sjson.SetRaw(fnRaw, "parametersJsonSchema.properties", `{}`)
 							if errSet != nil {
 								log.Warnf("Failed to set default schema properties for tool '%s': %v", fn.Get("name").String(), errSet)
 								continue
@@ -293,7 +324,7 @@ func ConvertOpenAIRequestToGeminiCLI(modelName string, inputRawJSON []byte, _ bo
 							log.Warnf("Failed to set default schema type for tool '%s': %v", fn.Get("name").String(), errSet)
 							continue
 						}
-						fnRaw, errSet = sjson.Set(fnRaw, "parametersJsonSchema.properties", map[string]interface{}{})
+						fnRaw, errSet = sjson.SetRaw(fnRaw, "parametersJsonSchema.properties", `{}`)
 						if errSet != nil {
 							log.Warnf("Failed to set default schema properties for tool '%s': %v", fn.Get("name").String(), errSet)
 							continue

internal/translator/gemini-cli/openai/chat-completions/gemini-cli_openai_response.go

@@ -8,7 +8,6 @@ package chat_completions
 import (
 	"bytes"
 	"context"
-	"encoding/json"
 	"fmt"
 	"strings"
 	"sync/atomic"
@@ -171,21 +170,16 @@ func ConvertCliResponseToOpenAI(_ context.Context, _ string, originalRequestRawJ
 					mimeType = "image/png"
 				}
 				imageURL := fmt.Sprintf("data:%s;base64,%s", mimeType, data)
-				imagePayload, err := json.Marshal(map[string]any{
-					"type": "image_url",
-					"image_url": map[string]string{
-						"url": imageURL,
-					},
-				})
-				if err != nil {
-					continue
-				}
 				imagesResult := gjson.Get(template, "choices.0.delta.images")
 				if !imagesResult.Exists() || !imagesResult.IsArray() {
 					template, _ = sjson.SetRaw(template, "choices.0.delta.images", `[]`)
 				}
+				imageIndex := len(gjson.Get(template, "choices.0.delta.images").Array())
+				imagePayload := `{"type":"image_url","image_url":{"url":""}}`
+				imagePayload, _ = sjson.Set(imagePayload, "index", imageIndex)
+				imagePayload, _ = sjson.Set(imagePayload, "image_url.url", imageURL)
 				template, _ = sjson.Set(template, "choices.0.delta.role", "assistant")
-				template, _ = sjson.SetRaw(template, "choices.0.delta.images.-1", string(imagePayload))
+				template, _ = sjson.SetRaw(template, "choices.0.delta.images.-1", imagePayload)
 			}
 		}
 	}

internal/translator/gemini/claude/gemini_claude_request.go

@@ -7,10 +7,8 @@ package claude
 
 import (
 	"bytes"
-	"encoding/json"
 	"strings"
 
-	client "github.com/router-for-me/CLIProxyAPI/v6/internal/interfaces"
 	"github.com/router-for-me/CLIProxyAPI/v6/internal/translator/gemini/common"
 	"github.com/router-for-me/CLIProxyAPI/v6/internal/util"
 	"github.com/tidwall/gjson"
@@ -34,92 +32,102 @@ func ConvertClaudeRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 	rawJSON := bytes.Clone(inputRawJSON)
 	rawJSON = bytes.Replace(rawJSON, []byte(`"url":{"type":"string","format":"uri",`), []byte(`"url":{"type":"string",`), -1)
 
+	// Build output Gemini CLI request JSON
+	out := `{"contents":[]}`
+	out, _ = sjson.Set(out, "model", modelName)
+
 	// system instruction
-	var systemInstruction *client.Content
-	systemResult := gjson.GetBytes(rawJSON, "system")
-	if systemResult.IsArray() {
-		systemResults := systemResult.Array()
-		systemInstruction = &client.Content{Role: "user", Parts: []client.Part{}}
-		for i := 0; i < len(systemResults); i++ {
-			systemPromptResult := systemResults[i]
-			systemTypePromptResult := systemPromptResult.Get("type")
-			if systemTypePromptResult.Type == gjson.String && systemTypePromptResult.String() == "text" {
-				systemPrompt := systemPromptResult.Get("text").String()
-				systemPart := client.Part{Text: systemPrompt}
-				systemInstruction.Parts = append(systemInstruction.Parts, systemPart)
+	if systemResult := gjson.GetBytes(rawJSON, "system"); systemResult.IsArray() {
+		systemInstruction := `{"role":"user","parts":[]}`
+		hasSystemParts := false
+		systemResult.ForEach(func(_, systemPromptResult gjson.Result) bool {
+			if systemPromptResult.Get("type").String() == "text" {
+				textResult := systemPromptResult.Get("text")
+				if textResult.Type == gjson.String {
+					part := `{"text":""}`
+					part, _ = sjson.Set(part, "text", textResult.String())
+					systemInstruction, _ = sjson.SetRaw(systemInstruction, "parts.-1", part)
+					hasSystemParts = true
+				}
 			}
+			return true
+		})
+		if hasSystemParts {
+			out, _ = sjson.SetRaw(out, "system_instruction", systemInstruction)
 		}
-		if len(systemInstruction.Parts) == 0 {
-			systemInstruction = nil
-		}
+	} else if systemResult.Type == gjson.String {
+		out, _ = sjson.Set(out, "system_instruction.parts.-1.text", systemResult.String())
 	}
 
 	// contents
-	contents := make([]client.Content, 0)
-	messagesResult := gjson.GetBytes(rawJSON, "messages")
-	if messagesResult.IsArray() {
-		messageResults := messagesResult.Array()
-		for i := 0; i < len(messageResults); i++ {
-			messageResult := messageResults[i]
+	if messagesResult := gjson.GetBytes(rawJSON, "messages"); messagesResult.IsArray() {
+		messagesResult.ForEach(func(_, messageResult gjson.Result) bool {
 			roleResult := messageResult.Get("role")
 			if roleResult.Type != gjson.String {
-				continue
+				return true
 			}
 			role := roleResult.String()
 			if role == "assistant" {
 				role = "model"
 			}
-			clientContent := client.Content{Role: role, Parts: []client.Part{}}
+
+			contentJSON := `{"role":"","parts":[]}`
+			contentJSON, _ = sjson.Set(contentJSON, "role", role)
+
 			contentsResult := messageResult.Get("content")
 			if contentsResult.IsArray() {
-				contentResults := contentsResult.Array()
-				for j := 0; j < len(contentResults); j++ {
-					contentResult := contentResults[j]
-					contentTypeResult := contentResult.Get("type")
-					if contentTypeResult.Type == gjson.String && contentTypeResult.String() == "text" {
-						prompt := contentResult.Get("text").String()
-						clientContent.Parts = append(clientContent.Parts, client.Part{Text: prompt})
-					} else if contentTypeResult.Type == gjson.String && contentTypeResult.String() == "tool_use" {
+				contentsResult.ForEach(func(_, contentResult gjson.Result) bool {
+					switch contentResult.Get("type").String() {
+					case "text":
+						part := `{"text":""}`
+						part, _ = sjson.Set(part, "text", contentResult.Get("text").String())
+						contentJSON, _ = sjson.SetRaw(contentJSON, "parts.-1", part)
+
+					case "tool_use":
 						functionName := contentResult.Get("name").String()
 						functionArgs := contentResult.Get("input").String()
-						var args map[string]any
-						if err := json.Unmarshal([]byte(functionArgs), &args); err == nil {
-							clientContent.Parts = append(clientContent.Parts, client.Part{
-								FunctionCall:     &client.FunctionCall{Name: functionName, Args: args},
-								ThoughtSignature: geminiClaudeThoughtSignature,
-							})
+						argsResult := gjson.Parse(functionArgs)
+						if argsResult.IsObject() && gjson.Valid(functionArgs) {
+							part := `{"thoughtSignature":"","functionCall":{"name":"","args":{}}}`
+							part, _ = sjson.Set(part, "thoughtSignature", geminiClaudeThoughtSignature)
+							part, _ = sjson.Set(part, "functionCall.name", functionName)
+							part, _ = sjson.SetRaw(part, "functionCall.args", functionArgs)
+							contentJSON, _ = sjson.SetRaw(contentJSON, "parts.-1", part)
 						}
-					} else if contentTypeResult.Type == gjson.String && contentTypeResult.String() == "tool_result" {
+
+					case "tool_result":
 						toolCallID := contentResult.Get("tool_use_id").String()
-						if toolCallID != "" {
-							funcName := toolCallID
-							toolCallIDs := strings.Split(toolCallID, "-")
-							if len(toolCallIDs) > 1 {
-								funcName = strings.Join(toolCallIDs[0:len(toolCallIDs)-1], "-")
-							}
-							responseData := contentResult.Get("content").Raw
-							functionResponse := client.FunctionResponse{Name: funcName, Response: map[string]interface{}{"result": responseData}}
-							clientContent.Parts = append(clientContent.Parts, client.Part{FunctionResponse: &functionResponse})
+						if toolCallID == "" {
+							return true
 						}
+						funcName := toolCallID
+						toolCallIDs := strings.Split(toolCallID, "-")
+						if len(toolCallIDs) > 1 {
+							funcName = strings.Join(toolCallIDs[0:len(toolCallIDs)-1], "-")
+						}
+						responseData := contentResult.Get("content").Raw
+						part := `{"functionResponse":{"name":"","response":{"result":""}}}`
+						part, _ = sjson.Set(part, "functionResponse.name", funcName)
+						part, _ = sjson.Set(part, "functionResponse.response.result", responseData)
+						contentJSON, _ = sjson.SetRaw(contentJSON, "parts.-1", part)
 					}
-				}
-				contents = append(contents, clientContent)
+					return true
+				})
+				out, _ = sjson.SetRaw(out, "contents.-1", contentJSON)
 			} else if contentsResult.Type == gjson.String {
-				prompt := contentsResult.String()
-				contents = append(contents, client.Content{Role: role, Parts: []client.Part{{Text: prompt}}})
+				part := `{"text":""}`
+				part, _ = sjson.Set(part, "text", contentsResult.String())
+				contentJSON, _ = sjson.SetRaw(contentJSON, "parts.-1", part)
+				out, _ = sjson.SetRaw(out, "contents.-1", contentJSON)
 			}
-		}
+			return true
+		})
 	}
 
 	// tools
-	var tools []client.ToolDeclaration
-	toolsResult := gjson.GetBytes(rawJSON, "tools")
-	if toolsResult.IsArray() {
-		tools = make([]client.ToolDeclaration, 1)
-		tools[0].FunctionDeclarations = make([]any, 0)
-		toolsResults := toolsResult.Array()
-		for i := 0; i < len(toolsResults); i++ {
-			toolResult := toolsResults[i]
+	if toolsResult := gjson.GetBytes(rawJSON, "tools"); toolsResult.IsArray() {
+		hasTools := false
+		toolsResult.ForEach(func(_, toolResult gjson.Result) bool {
 			inputSchemaResult := toolResult.Get("input_schema")
 			if inputSchemaResult.Exists() && inputSchemaResult.IsObject() {
 				inputSchema := inputSchemaResult.Raw
@@ -129,30 +137,19 @@ func ConvertClaudeRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 				tool, _ = sjson.Delete(tool, "input_examples")
 				tool, _ = sjson.Delete(tool, "type")
 				tool, _ = sjson.Delete(tool, "cache_control")
-				var toolDeclaration any
-				if err := json.Unmarshal([]byte(tool), &toolDeclaration); err == nil {
-					tools[0].FunctionDeclarations = append(tools[0].FunctionDeclarations, toolDeclaration)
+				if gjson.Valid(tool) && gjson.Parse(tool).IsObject() {
+					if !hasTools {
+						out, _ = sjson.SetRaw(out, "tools", `[{"functionDeclarations":[]}]`)
+						hasTools = true
+					}
+					out, _ = sjson.SetRaw(out, "tools.0.functionDeclarations.-1", tool)
 				}
 			}
+			return true
+		})
+		if !hasTools {
+			out, _ = sjson.Delete(out, "tools")
 		}
-	} else {
-		tools = make([]client.ToolDeclaration, 0)
-	}
-
-	// Build output Gemini CLI request JSON
-	out := `{"contents":[]}`
-	out, _ = sjson.Set(out, "model", modelName)
-	if systemInstruction != nil {
-		b, _ := json.Marshal(systemInstruction)
-		out, _ = sjson.SetRaw(out, "system_instruction", string(b))
-	}
-	if len(contents) > 0 {
-		b, _ := json.Marshal(contents)
-		out, _ = sjson.SetRaw(out, "contents", string(b))
-	}
-	if len(tools) > 0 && len(tools[0].FunctionDeclarations) > 0 {
-		b, _ := json.Marshal(tools)
-		out, _ = sjson.SetRaw(out, "tools", string(b))
 	}
 
 	// Map Anthropic thinking -> Gemini thinkingBudget/include_thoughts when enabled

internal/translator/gemini/claude/gemini_claude_response.go

@@ -9,7 +9,6 @@ package claude
 import (
 	"bytes"
 	"context"
-	"encoding/json"
 	"fmt"
 	"strings"
 	"sync/atomic"
@@ -282,22 +281,16 @@ func ConvertGeminiResponseToClaudeNonStream(_ context.Context, _ string, origina
 
 	root := gjson.ParseBytes(rawJSON)
 
-	response := map[string]interface{}{
-		"id":            root.Get("responseId").String(),
-		"type":          "message",
-		"role":          "assistant",
-		"model":         root.Get("modelVersion").String(),
-		"content":       []interface{}{},
-		"stop_reason":   nil,
-		"stop_sequence": nil,
-		"usage": map[string]interface{}{
-			"input_tokens":  root.Get("usageMetadata.promptTokenCount").Int(),
-			"output_tokens": root.Get("usageMetadata.candidatesTokenCount").Int() + root.Get("usageMetadata.thoughtsTokenCount").Int(),
-		},
-	}
+	out := `{"id":"","type":"message","role":"assistant","model":"","content":[],"stop_reason":null,"stop_sequence":null,"usage":{"input_tokens":0,"output_tokens":0}}`
+	out, _ = sjson.Set(out, "id", root.Get("responseId").String())
+	out, _ = sjson.Set(out, "model", root.Get("modelVersion").String())
+
+	inputTokens := root.Get("usageMetadata.promptTokenCount").Int()
+	outputTokens := root.Get("usageMetadata.candidatesTokenCount").Int() + root.Get("usageMetadata.thoughtsTokenCount").Int()
+	out, _ = sjson.Set(out, "usage.input_tokens", inputTokens)
+	out, _ = sjson.Set(out, "usage.output_tokens", outputTokens)
 
 	parts := root.Get("candidates.0.content.parts")
-	var contentBlocks []interface{}
 	textBuilder := strings.Builder{}
 	thinkingBuilder := strings.Builder{}
 	toolIDCounter := 0
@@ -307,10 +300,9 @@ func ConvertGeminiResponseToClaudeNonStream(_ context.Context, _ string, origina
 		if textBuilder.Len() == 0 {
 			return
 		}
-		contentBlocks = append(contentBlocks, map[string]interface{}{
-			"type": "text",
-			"text": textBuilder.String(),
-		})
+		block := `{"type":"text","text":""}`
+		block, _ = sjson.Set(block, "text", textBuilder.String())
+		out, _ = sjson.SetRaw(out, "content.-1", block)
 		textBuilder.Reset()
 	}
 
@@ -318,10 +310,9 @@ func ConvertGeminiResponseToClaudeNonStream(_ context.Context, _ string, origina
 		if thinkingBuilder.Len() == 0 {
 			return
 		}
-		contentBlocks = append(contentBlocks, map[string]interface{}{
-			"type":     "thinking",
-			"thinking": thinkingBuilder.String(),
-		})
+		block := `{"type":"thinking","thinking":""}`
+		block, _ = sjson.Set(block, "thinking", thinkingBuilder.String())
+		out, _ = sjson.SetRaw(out, "content.-1", block)
 		thinkingBuilder.Reset()
 	}
 
@@ -345,21 +336,15 @@ func ConvertGeminiResponseToClaudeNonStream(_ context.Context, _ string, origina
 
 				name := functionCall.Get("name").String()
 				toolIDCounter++
-				toolBlock := map[string]interface{}{
-					"type":  "tool_use",
-					"id":    fmt.Sprintf("tool_%d", toolIDCounter),
-					"name":  name,
-					"input": map[string]interface{}{},
+				toolBlock := `{"type":"tool_use","id":"","name":"","input":{}}`
+				toolBlock, _ = sjson.Set(toolBlock, "id", fmt.Sprintf("tool_%d", toolIDCounter))
+				toolBlock, _ = sjson.Set(toolBlock, "name", name)
+				inputRaw := "{}"
+				if args := functionCall.Get("args"); args.Exists() && gjson.Valid(args.Raw) && args.IsObject() {
+					inputRaw = args.Raw
 				}
-
-				if args := functionCall.Get("args"); args.Exists() {
-					var parsed interface{}
-					if err := json.Unmarshal([]byte(args.Raw), &parsed); err == nil {
-						toolBlock["input"] = parsed
-					}
-				}
-
-				contentBlocks = append(contentBlocks, toolBlock)
+				toolBlock, _ = sjson.SetRaw(toolBlock, "input", inputRaw)
+				out, _ = sjson.SetRaw(out, "content.-1", toolBlock)
 				continue
 			}
 		}
@@ -368,8 +353,6 @@ func ConvertGeminiResponseToClaudeNonStream(_ context.Context, _ string, origina
 	flushThinking()
 	flushText()
 
-	response["content"] = contentBlocks
-
 	stopReason := "end_turn"
 	if hasToolCall {
 		stopReason = "tool_use"
@@ -385,19 +368,13 @@ func ConvertGeminiResponseToClaudeNonStream(_ context.Context, _ string, origina
 			}
 		}
 	}
-	response["stop_reason"] = stopReason
+	out, _ = sjson.Set(out, "stop_reason", stopReason)
 
-	if usage := response["usage"].(map[string]interface{}); usage["input_tokens"] == int64(0) && usage["output_tokens"] == int64(0) {
-		if usageMeta := root.Get("usageMetadata"); !usageMeta.Exists() {
-			delete(response, "usage")
-		}
+	if inputTokens == int64(0) && outputTokens == int64(0) && !root.Get("usageMetadata").Exists() {
+		out, _ = sjson.Delete(out, "usage")
 	}
 
-	encoded, err := json.Marshal(response)
-	if err != nil {
-		return ""
-	}
-	return string(encoded)
+	return out
 }
 
 func ClaudeTokenCount(ctx context.Context, count int64) string {

internal/translator/gemini/openai/chat-completions/gemini_openai_request.go

@@ -178,6 +178,14 @@ func ConvertOpenAIRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 				} else if content.IsObject() && content.Get("type").String() == "text" {
 					out, _ = sjson.SetBytes(out, "system_instruction.role", "user")
 					out, _ = sjson.SetBytes(out, "system_instruction.parts.0.text", content.Get("text").String())
+				} else if content.IsArray() {
+					contents := content.Array()
+					if len(contents) > 0 {
+						out, _ = sjson.SetBytes(out, "request.systemInstruction.role", "user")
+						for j := 0; j < len(contents); j++ {
+							out, _ = sjson.SetBytes(out, fmt.Sprintf("request.systemInstruction.parts.%d.text", j), contents[j].Get("text").String())
+						}
+					}
 				}
 			} else if role == "user" || (role == "system" && len(arr) == 1) {
 				// Build single user content node to avoid splitting into multiple contents
@@ -225,18 +233,15 @@ func ConvertOpenAIRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 			} else if role == "assistant" {
 				node := []byte(`{"role":"model","parts":[]}`)
 				p := 0
-
 				if content.Type == gjson.String {
 					// Assistant text -> single model content
 					node, _ = sjson.SetBytes(node, "parts.-1.text", content.String())
-					out, _ = sjson.SetRawBytes(out, "contents.-1", node)
 					p++
 				} else if content.IsArray() {
 					// Assistant multimodal content (e.g. text + image) -> single model content with parts
 					for _, item := range content.Array() {
 						switch item.Get("type").String() {
 						case "text":
-							node, _ = sjson.SetBytes(node, "parts."+itoa(p)+".text", item.Get("text").String())
 							p++
 						case "image_url":
 							// If the assistant returned an inline data URL, preserve it for history fidelity.
@@ -253,7 +258,6 @@ func ConvertOpenAIRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 							}
 						}
 					}
-					out, _ = sjson.SetRawBytes(out, "contents.-1", node)
 				}
 
 				// Tool calls -> single model content with functionCall parts
@@ -278,7 +282,7 @@ func ConvertOpenAIRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 					out, _ = sjson.SetRawBytes(out, "contents.-1", node)
 
 					// Append a single tool content combining name + response per function
-					toolNode := []byte(`{"role":"tool","parts":[]}`)
+					toolNode := []byte(`{"role":"user","parts":[]}`)
 					pp := 0
 					for _, fid := range fIDs {
 						if name, ok := tcID2Name[fid]; ok {
@@ -294,6 +298,8 @@ func ConvertOpenAIRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 					if pp > 0 {
 						out, _ = sjson.SetRawBytes(out, "contents.-1", toolNode)
 					}
+				} else {
+					out, _ = sjson.SetRawBytes(out, "contents.-1", node)
 				}
 			}
 		}
@@ -320,7 +326,7 @@ func ConvertOpenAIRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 								log.Warnf("Failed to set default schema type for tool '%s': %v", fn.Get("name").String(), errSet)
 								continue
 							}
-							fnRaw, errSet = sjson.Set(fnRaw, "parametersJsonSchema.properties", map[string]interface{}{})
+							fnRaw, errSet = sjson.SetRaw(fnRaw, "parametersJsonSchema.properties", `{}`)
 							if errSet != nil {
 								log.Warnf("Failed to set default schema properties for tool '%s': %v", fn.Get("name").String(), errSet)
 								continue
@@ -335,7 +341,7 @@ func ConvertOpenAIRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 							log.Warnf("Failed to set default schema type for tool '%s': %v", fn.Get("name").String(), errSet)
 							continue
 						}
-						fnRaw, errSet = sjson.Set(fnRaw, "parametersJsonSchema.properties", map[string]interface{}{})
+						fnRaw, errSet = sjson.SetRaw(fnRaw, "parametersJsonSchema.properties", `{}`)
 						if errSet != nil {
 							log.Warnf("Failed to set default schema properties for tool '%s': %v", fn.Get("name").String(), errSet)
 							continue

internal/translator/gemini/openai/chat-completions/gemini_openai_response.go

@@ -8,12 +8,12 @@ package chat_completions
 import (
 	"bytes"
 	"context"
-	"encoding/json"
 	"fmt"
 	"strings"
 	"sync/atomic"
 	"time"
 
+	log "github.com/sirupsen/logrus"
 	"github.com/tidwall/gjson"
 	"github.com/tidwall/sjson"
 )
@@ -89,18 +89,27 @@ func ConvertGeminiResponseToOpenAI(_ context.Context, _ string, originalRequestR
 
 	// Extract and set usage metadata (token counts).
 	if usageResult := gjson.GetBytes(rawJSON, "usageMetadata"); usageResult.Exists() {
+		cachedTokenCount := usageResult.Get("cachedContentTokenCount").Int()
 		if candidatesTokenCountResult := usageResult.Get("candidatesTokenCount"); candidatesTokenCountResult.Exists() {
 			template, _ = sjson.Set(template, "usage.completion_tokens", candidatesTokenCountResult.Int())
 		}
 		if totalTokenCountResult := usageResult.Get("totalTokenCount"); totalTokenCountResult.Exists() {
 			template, _ = sjson.Set(template, "usage.total_tokens", totalTokenCountResult.Int())
 		}
-		promptTokenCount := usageResult.Get("promptTokenCount").Int()
+		promptTokenCount := usageResult.Get("promptTokenCount").Int() - cachedTokenCount
 		thoughtsTokenCount := usageResult.Get("thoughtsTokenCount").Int()
 		template, _ = sjson.Set(template, "usage.prompt_tokens", promptTokenCount+thoughtsTokenCount)
 		if thoughtsTokenCount > 0 {
 			template, _ = sjson.Set(template, "usage.completion_tokens_details.reasoning_tokens", thoughtsTokenCount)
 		}
+		// Include cached token count if present (indicates prompt caching is working)
+		if cachedTokenCount > 0 {
+			var err error
+			template, err = sjson.Set(template, "usage.prompt_tokens_details.cached_tokens", cachedTokenCount)
+			if err != nil {
+				log.Warnf("gemini openai response: failed to set cached_tokens in streaming: %v", err)
+			}
+		}
 	}
 
 	// Process the main content part of the response.
@@ -173,21 +182,16 @@ func ConvertGeminiResponseToOpenAI(_ context.Context, _ string, originalRequestR
 					mimeType = "image/png"
 				}
 				imageURL := fmt.Sprintf("data:%s;base64,%s", mimeType, data)
-				imagePayload, err := json.Marshal(map[string]any{
-					"type": "image_url",
-					"image_url": map[string]string{
-						"url": imageURL,
-					},
-				})
-				if err != nil {
-					continue
-				}
 				imagesResult := gjson.Get(template, "choices.0.delta.images")
 				if !imagesResult.Exists() || !imagesResult.IsArray() {
 					template, _ = sjson.SetRaw(template, "choices.0.delta.images", `[]`)
 				}
+				imageIndex := len(gjson.Get(template, "choices.0.delta.images").Array())
+				imagePayload := `{"type":"image_url","image_url":{"url":""}}`
+				imagePayload, _ = sjson.Set(imagePayload, "index", imageIndex)
+				imagePayload, _ = sjson.Set(imagePayload, "image_url.url", imageURL)
 				template, _ = sjson.Set(template, "choices.0.delta.role", "assistant")
-				template, _ = sjson.SetRaw(template, "choices.0.delta.images.-1", string(imagePayload))
+				template, _ = sjson.SetRaw(template, "choices.0.delta.images.-1", imagePayload)
 			}
 		}
 	}
@@ -248,10 +252,19 @@ func ConvertGeminiResponseToOpenAINonStream(_ context.Context, _ string, origina
 		}
 		promptTokenCount := usageResult.Get("promptTokenCount").Int()
 		thoughtsTokenCount := usageResult.Get("thoughtsTokenCount").Int()
+		cachedTokenCount := usageResult.Get("cachedContentTokenCount").Int()
 		template, _ = sjson.Set(template, "usage.prompt_tokens", promptTokenCount+thoughtsTokenCount)
 		if thoughtsTokenCount > 0 {
 			template, _ = sjson.Set(template, "usage.completion_tokens_details.reasoning_tokens", thoughtsTokenCount)
 		}
+		// Include cached token count if present (indicates prompt caching is working)
+		if cachedTokenCount > 0 {
+			var err error
+			template, err = sjson.Set(template, "usage.prompt_tokens_details.cached_tokens", cachedTokenCount)
+			if err != nil {
+				log.Warnf("gemini openai response: failed to set cached_tokens in non-streaming: %v", err)
+			}
+		}
 	}
 
 	// Process the main content part of the response.
@@ -305,21 +318,16 @@ func ConvertGeminiResponseToOpenAINonStream(_ context.Context, _ string, origina
 					mimeType = "image/png"
 				}
 				imageURL := fmt.Sprintf("data:%s;base64,%s", mimeType, data)
-				imagePayload, err := json.Marshal(map[string]any{
-					"type": "image_url",
-					"image_url": map[string]string{
-						"url": imageURL,
-					},
-				})
-				if err != nil {
-					continue
-				}
 				imagesResult := gjson.Get(template, "choices.0.message.images")
 				if !imagesResult.Exists() || !imagesResult.IsArray() {
 					template, _ = sjson.SetRaw(template, "choices.0.message.images", `[]`)
 				}
+				imageIndex := len(gjson.Get(template, "choices.0.message.images").Array())
+				imagePayload := `{"type":"image_url","image_url":{"url":""}}`
+				imagePayload, _ = sjson.Set(imagePayload, "index", imageIndex)
+				imagePayload, _ = sjson.Set(imagePayload, "image_url.url", imageURL)
 				template, _ = sjson.Set(template, "choices.0.message.role", "assistant")
-				template, _ = sjson.SetRaw(template, "choices.0.message.images.-1", string(imagePayload))
+				template, _ = sjson.SetRaw(template, "choices.0.message.images.-1", imagePayload)
 			}
 		}
 	}

internal/translator/gemini/openai/responses/gemini_openai-responses_response.go

@@ -23,6 +23,7 @@ type geminiToResponsesState struct {
 	MsgIndex     int
 	CurrentMsgID string
 	TextBuf      strings.Builder
+	ItemTextBuf  strings.Builder
 
 	// reasoning aggregation
 	ReasoningOpened bool
@@ -189,6 +190,8 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 					partAdded, _ = sjson.Set(partAdded, "item_id", st.CurrentMsgID)
 					partAdded, _ = sjson.Set(partAdded, "output_index", st.MsgIndex)
 					out = append(out, emitEvent("response.content_part.added", partAdded))
+					st.ItemTextBuf.Reset()
+					st.ItemTextBuf.WriteString(t.String())
 				}
 				st.TextBuf.WriteString(t.String())
 				msg := `{"type":"response.output_text.delta","sequence_number":0,"item_id":"","output_index":0,"content_index":0,"delta":"","logprobs":[]}`
@@ -250,20 +253,24 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 		finalizeReasoning()
 		// Close message output if opened
 		if st.MsgOpened {
+			fullText := st.ItemTextBuf.String()
 			done := `{"type":"response.output_text.done","sequence_number":0,"item_id":"","output_index":0,"content_index":0,"text":"","logprobs":[]}`
 			done, _ = sjson.Set(done, "sequence_number", nextSeq())
 			done, _ = sjson.Set(done, "item_id", st.CurrentMsgID)
 			done, _ = sjson.Set(done, "output_index", st.MsgIndex)
+			done, _ = sjson.Set(done, "text", fullText)
 			out = append(out, emitEvent("response.output_text.done", done))
 			partDone := `{"type":"response.content_part.done","sequence_number":0,"item_id":"","output_index":0,"content_index":0,"part":{"type":"output_text","annotations":[],"logprobs":[],"text":""}}`
 			partDone, _ = sjson.Set(partDone, "sequence_number", nextSeq())
 			partDone, _ = sjson.Set(partDone, "item_id", st.CurrentMsgID)
 			partDone, _ = sjson.Set(partDone, "output_index", st.MsgIndex)
+			partDone, _ = sjson.Set(partDone, "part.text", fullText)
 			out = append(out, emitEvent("response.content_part.done", partDone))
 			final := `{"type":"response.output_item.done","sequence_number":0,"output_index":0,"item":{"id":"","type":"message","status":"completed","content":[{"type":"output_text","text":""}],"role":"assistant"}}`
 			final, _ = sjson.Set(final, "sequence_number", nextSeq())
 			final, _ = sjson.Set(final, "output_index", st.MsgIndex)
 			final, _ = sjson.Set(final, "item.id", st.CurrentMsgID)
+			final, _ = sjson.Set(final, "item.content.0.text", fullText)
 			out = append(out, emitEvent("response.output_item.done", final))
 		}
 
@@ -377,27 +384,18 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 		}
 
 		// Compose outputs in encountered order: reasoning, message, function_calls
-		var outputs []interface{}
+		outputsWrapper := `{"arr":[]}`
 		if st.ReasoningOpened {
-			outputs = append(outputs, map[string]interface{}{
-				"id":      st.ReasoningItemID,
-				"type":    "reasoning",
-				"summary": []interface{}{map[string]interface{}{"type": "summary_text", "text": st.ReasoningBuf.String()}},
-			})
+			item := `{"id":"","type":"reasoning","summary":[{"type":"summary_text","text":""}]}`
+			item, _ = sjson.Set(item, "id", st.ReasoningItemID)
+			item, _ = sjson.Set(item, "summary.0.text", st.ReasoningBuf.String())
+			outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 		}
 		if st.MsgOpened {
-			outputs = append(outputs, map[string]interface{}{
-				"id":     st.CurrentMsgID,
-				"type":   "message",
-				"status": "completed",
-				"content": []interface{}{map[string]interface{}{
-					"type":        "output_text",
-					"annotations": []interface{}{},
-					"logprobs":    []interface{}{},
-					"text":        st.TextBuf.String(),
-				}},
-				"role": "assistant",
-			})
+			item := `{"id":"","type":"message","status":"completed","content":[{"type":"output_text","annotations":[],"logprobs":[],"text":""}],"role":"assistant"}`
+			item, _ = sjson.Set(item, "id", st.CurrentMsgID)
+			item, _ = sjson.Set(item, "content.0.text", st.TextBuf.String())
+			outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 		}
 		if len(st.FuncArgsBuf) > 0 {
 			idxs := make([]int, 0, len(st.FuncArgsBuf))
@@ -416,18 +414,16 @@ func ConvertGeminiResponseToOpenAIResponses(_ context.Context, modelName string,
 				if b := st.FuncArgsBuf[idx]; b != nil {
 					args = b.String()
 				}
-				outputs = append(outputs, map[string]interface{}{
-					"id":        fmt.Sprintf("fc_%s", st.FuncCallIDs[idx]),
-					"type":      "function_call",
-					"status":    "completed",
-					"arguments": args,
-					"call_id":   st.FuncCallIDs[idx],
-					"name":      st.FuncNames[idx],
-				})
+				item := `{"id":"","type":"function_call","status":"completed","arguments":"","call_id":"","name":""}`
+				item, _ = sjson.Set(item, "id", fmt.Sprintf("fc_%s", st.FuncCallIDs[idx]))
+				item, _ = sjson.Set(item, "arguments", args)
+				item, _ = sjson.Set(item, "call_id", st.FuncCallIDs[idx])
+				item, _ = sjson.Set(item, "name", st.FuncNames[idx])
+				outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 			}
 		}
-		if len(outputs) > 0 {
-			completed, _ = sjson.Set(completed, "response.output", outputs)
+		if gjson.Get(outputsWrapper, "arr.#").Int() > 0 {
+			completed, _ = sjson.SetRaw(completed, "response.output", gjson.Get(outputsWrapper, "arr").Raw)
 		}
 
 		// usage mapping
@@ -558,11 +554,24 @@ func ConvertGeminiResponseToOpenAIResponsesNonStream(_ context.Context, _ string
 	}
 
 	// Build outputs from candidates[0].content.parts
-	var outputs []interface{}
 	var reasoningText strings.Builder
 	var reasoningEncrypted string
 	var messageText strings.Builder
 	var haveMessage bool
+
+	haveOutput := false
+	ensureOutput := func() {
+		if haveOutput {
+			return
+		}
+		resp, _ = sjson.SetRaw(resp, "output", "[]")
+		haveOutput = true
+	}
+	appendOutput := func(itemJSON string) {
+		ensureOutput()
+		resp, _ = sjson.SetRaw(resp, "output.-1", itemJSON)
+	}
+
 	if parts := root.Get("candidates.0.content.parts"); parts.Exists() && parts.IsArray() {
 		parts.ForEach(func(_, p gjson.Result) bool {
 			if p.Get("thought").Bool() {
@@ -583,19 +592,16 @@ func ConvertGeminiResponseToOpenAIResponsesNonStream(_ context.Context, _ string
 				name := fc.Get("name").String()
 				args := fc.Get("args")
 				callID := fmt.Sprintf("call_%x_%d", time.Now().UnixNano(), atomic.AddUint64(&funcCallIDCounter, 1))
-				outputs = append(outputs, map[string]interface{}{
-					"id":     fmt.Sprintf("fc_%s", callID),
-					"type":   "function_call",
-					"status": "completed",
-					"arguments": func() string {
-						if args.Exists() {
-							return args.Raw
-						}
-						return ""
-					}(),
-					"call_id": callID,
-					"name":    name,
-				})
+				itemJSON := `{"id":"","type":"function_call","status":"completed","arguments":"","call_id":"","name":""}`
+				itemJSON, _ = sjson.Set(itemJSON, "id", fmt.Sprintf("fc_%s", callID))
+				itemJSON, _ = sjson.Set(itemJSON, "call_id", callID)
+				itemJSON, _ = sjson.Set(itemJSON, "name", name)
+				argsStr := ""
+				if args.Exists() {
+					argsStr = args.Raw
+				}
+				itemJSON, _ = sjson.Set(itemJSON, "arguments", argsStr)
+				appendOutput(itemJSON)
 				return true
 			}
 			return true
@@ -605,42 +611,24 @@ func ConvertGeminiResponseToOpenAIResponsesNonStream(_ context.Context, _ string
 	// Reasoning output item
 	if reasoningText.Len() > 0 || reasoningEncrypted != "" {
 		rid := strings.TrimPrefix(id, "resp_")
-		item := map[string]interface{}{
-			"id":                fmt.Sprintf("rs_%s", rid),
-			"type":              "reasoning",
-			"encrypted_content": reasoningEncrypted,
-		}
-		var summaries []interface{}
+		itemJSON := `{"id":"","type":"reasoning","encrypted_content":""}`
+		itemJSON, _ = sjson.Set(itemJSON, "id", fmt.Sprintf("rs_%s", rid))
+		itemJSON, _ = sjson.Set(itemJSON, "encrypted_content", reasoningEncrypted)
 		if reasoningText.Len() > 0 {
-			summaries = append(summaries, map[string]interface{}{
-				"type": "summary_text",
-				"text": reasoningText.String(),
-			})
+			summaryJSON := `{"type":"summary_text","text":""}`
+			summaryJSON, _ = sjson.Set(summaryJSON, "text", reasoningText.String())
+			itemJSON, _ = sjson.SetRaw(itemJSON, "summary", "[]")
+			itemJSON, _ = sjson.SetRaw(itemJSON, "summary.-1", summaryJSON)
 		}
-		if summaries != nil {
-			item["summary"] = summaries
-		}
-		outputs = append(outputs, item)
+		appendOutput(itemJSON)
 	}
 
 	// Assistant message output item
 	if haveMessage {
-		outputs = append(outputs, map[string]interface{}{
-			"id":     fmt.Sprintf("msg_%s_0", strings.TrimPrefix(id, "resp_")),
-			"type":   "message",
-			"status": "completed",
-			"content": []interface{}{map[string]interface{}{
-				"type":        "output_text",
-				"annotations": []interface{}{},
-				"logprobs":    []interface{}{},
-				"text":        messageText.String(),
-			}},
-			"role": "assistant",
-		})
-	}
-
-	if len(outputs) > 0 {
-		resp, _ = sjson.Set(resp, "output", outputs)
+		itemJSON := `{"id":"","type":"message","status":"completed","content":[{"type":"output_text","annotations":[],"logprobs":[],"text":""}],"role":"assistant"}`
+		itemJSON, _ = sjson.Set(itemJSON, "id", fmt.Sprintf("msg_%s_0", strings.TrimPrefix(id, "resp_")))
+		itemJSON, _ = sjson.Set(itemJSON, "content.0.text", messageText.String())
+		appendOutput(itemJSON)
 	}
 
 	// usage mapping

internal/translator/kiro/common/message_merge.go

@@ -10,6 +10,7 @@ import (
 // MergeAdjacentMessages merges adjacent messages with the same role.
 // This reduces API call complexity and improves compatibility.
 // Based on AIClient-2-API implementation.
+// NOTE: Tool messages are NOT merged because each has a unique tool_call_id that must be preserved.
 func MergeAdjacentMessages(messages []gjson.Result) []gjson.Result {
 	if len(messages) <= 1 {
 		return messages
@@ -26,6 +27,12 @@ func MergeAdjacentMessages(messages []gjson.Result) []gjson.Result {
 		currentRole := msg.Get("role").String()
 		lastRole := lastMsg.Get("role").String()
 
+		// Don't merge tool messages - each has a unique tool_call_id
+		if currentRole == "tool" || lastRole == "tool" {
+			merged = append(merged, msg)
+			continue
+		}
+
 		if currentRole == lastRole {
 			// Merge content from current message into last message
 			mergedContent := mergeMessageContent(lastMsg, msg)

internal/translator/kiro/openai/kiro_openai_request.go

@@ -450,24 +450,10 @@ func processOpenAIMessages(messages gjson.Result, modelID, origin string) ([]Kir
 	// Merge adjacent messages with the same role
 	messagesArray := kirocommon.MergeAdjacentMessages(messages.Array())
 
-	// Build tool_call_id to name mapping from assistant messages
-	toolCallIDToName := make(map[string]string)
-	for _, msg := range messagesArray {
-		if msg.Get("role").String() == "assistant" {
-			toolCalls := msg.Get("tool_calls")
-			if toolCalls.IsArray() {
-				for _, tc := range toolCalls.Array() {
-					if tc.Get("type").String() == "function" {
-						id := tc.Get("id").String()
-						name := tc.Get("function.name").String()
-						if id != "" && name != "" {
-							toolCallIDToName[id] = name
-						}
-					}
-				}
-			}
-		}
-	}
+	// Track pending tool results that should be attached to the next user message
+	// This is critical for LiteLLM-translated requests where tool results appear
+	// as separate "tool" role messages between assistant and user messages
+	var pendingToolResults []KiroToolResult
 
 	for i, msg := range messagesArray {
 		role := msg.Get("role").String()
@@ -480,6 +466,10 @@ func processOpenAIMessages(messages gjson.Result, modelID, origin string) ([]Kir
 
 		case "user":
 			userMsg, toolResults := buildUserMessageFromOpenAI(msg, modelID, origin)
+			// Merge any pending tool results from preceding "tool" role messages
+			toolResults = append(pendingToolResults, toolResults...)
+			pendingToolResults = nil // Reset pending tool results
+
 			if isLastMessage {
 				currentUserMsg = &userMsg
 				currentToolResults = toolResults
@@ -505,6 +495,24 @@ func processOpenAIMessages(messages gjson.Result, modelID, origin string) ([]Kir
 
 		case "assistant":
 			assistantMsg := buildAssistantMessageFromOpenAI(msg)
+
+			// If there are pending tool results, we need to insert a synthetic user message
+			// before this assistant message to maintain proper conversation structure
+			if len(pendingToolResults) > 0 {
+				syntheticUserMsg := KiroUserInputMessage{
+					Content: "Tool results provided.",
+					ModelID: modelID,
+					Origin:  origin,
+					UserInputMessageContext: &KiroUserInputMessageContext{
+						ToolResults: pendingToolResults,
+					},
+				}
+				history = append(history, KiroHistoryMessage{
+					UserInputMessage: &syntheticUserMsg,
+				})
+				pendingToolResults = nil
+			}
+
 			if isLastMessage {
 				history = append(history, KiroHistoryMessage{
 					AssistantResponseMessage: &assistantMsg,
@@ -524,7 +532,7 @@ func processOpenAIMessages(messages gjson.Result, modelID, origin string) ([]Kir
 		case "tool":
 			// Tool messages in OpenAI format provide results for tool_calls
 			// These are typically followed by user or assistant messages
-			// Process them and merge into the next user message's tool results
+			// Collect them as pending and attach to the next user message
 			toolCallID := msg.Get("tool_call_id").String()
 			content := msg.Get("content").String()
 
@@ -534,9 +542,21 @@ func processOpenAIMessages(messages gjson.Result, modelID, origin string) ([]Kir
 					Content:   []KiroTextContent{{Text: content}},
 					Status:    "success",
 				}
-				// Tool results should be included in the next user message
-				// For now, collect them and they'll be handled when we build the current message
-				currentToolResults = append(currentToolResults, toolResult)
+				// Collect pending tool results to attach to the next user message
+				pendingToolResults = append(pendingToolResults, toolResult)
+			}
+		}
+	}
+
+	// Handle case where tool results are at the end with no following user message
+	if len(pendingToolResults) > 0 {
+		currentToolResults = append(currentToolResults, pendingToolResults...)
+		// If there's no current user message, create a synthetic one for the tool results
+		if currentUserMsg == nil {
+			currentUserMsg = &KiroUserInputMessage{
+				Content: "Tool results provided.",
+				ModelID: modelID,
+				Origin:  origin,
 			}
 		}
 	}
@@ -551,9 +571,6 @@ func buildUserMessageFromOpenAI(msg gjson.Result, modelID, origin string) (KiroU
 	var toolResults []KiroToolResult
 	var images []KiroImage
 
-	// Track seen toolCallIds to deduplicate
-	seenToolCallIDs := make(map[string]bool)
-
 	if content.IsArray() {
 		for _, part := range content.Array() {
 			partType := part.Get("type").String()
@@ -589,9 +606,6 @@ func buildUserMessageFromOpenAI(msg gjson.Result, modelID, origin string) (KiroU
 		contentBuilder.WriteString(content.String())
 	}
 
-	// Check for tool_calls in the message (shouldn't be in user messages, but handle edge cases)
-	_ = seenToolCallIDs // Used for deduplication if needed
-
 	userMsg := KiroUserInputMessage{
 		Content: contentBuilder.String(),
 		ModelID: modelID,

internal/translator/kiro/openai/kiro_openai_request_test.go

@@ -0,0 +1,386 @@
+package openai
+
+import (
+	"encoding/json"
+	"testing"
+)
+
+// TestToolResultsAttachedToCurrentMessage verifies that tool results from "tool" role messages
+// are properly attached to the current user message (the last message in the conversation).
+// This is critical for LiteLLM-translated requests where tool results appear as separate messages.
+func TestToolResultsAttachedToCurrentMessage(t *testing.T) {
+	// OpenAI format request simulating LiteLLM's translation from Anthropic format
+	// Sequence: user -> assistant (with tool_calls) -> tool (result) -> user
+	// The last user message should have the tool results attached
+	input := []byte(`{
+		"model": "kiro-claude-opus-4-5-agentic",
+		"messages": [
+			{"role": "user", "content": "Hello, can you read a file for me?"},
+			{
+				"role": "assistant",
+				"content": "I'll read that file for you.",
+				"tool_calls": [
+					{
+						"id": "call_abc123",
+						"type": "function",
+						"function": {
+							"name": "Read",
+							"arguments": "{\"file_path\": \"/tmp/test.txt\"}"
+						}
+					}
+				]
+			},
+			{
+				"role": "tool",
+				"tool_call_id": "call_abc123",
+				"content": "File contents: Hello World!"
+			},
+			{"role": "user", "content": "What did the file say?"}
+		]
+	}`)
+
+	result, _ := BuildKiroPayloadFromOpenAI(input, "kiro-model", "", "CLI", false, false, nil, nil)
+
+	var payload KiroPayload
+	if err := json.Unmarshal(result, &payload); err != nil {
+		t.Fatalf("Failed to unmarshal result: %v", err)
+	}
+
+	// The last user message becomes currentMessage
+	// History should have: user (first), assistant (with tool_calls)
+	t.Logf("History count: %d", len(payload.ConversationState.History))
+	if len(payload.ConversationState.History) != 2 {
+		t.Errorf("Expected 2 history entries (user + assistant), got %d", len(payload.ConversationState.History))
+	}
+
+	// Tool results should be attached to currentMessage (the last user message)
+	ctx := payload.ConversationState.CurrentMessage.UserInputMessage.UserInputMessageContext
+	if ctx == nil {
+		t.Fatal("Expected currentMessage to have UserInputMessageContext with tool results")
+	}
+
+	if len(ctx.ToolResults) != 1 {
+		t.Fatalf("Expected 1 tool result in currentMessage, got %d", len(ctx.ToolResults))
+	}
+
+	tr := ctx.ToolResults[0]
+	if tr.ToolUseID != "call_abc123" {
+		t.Errorf("Expected toolUseId 'call_abc123', got '%s'", tr.ToolUseID)
+	}
+	if len(tr.Content) == 0 || tr.Content[0].Text != "File contents: Hello World!" {
+		t.Errorf("Tool result content mismatch, got: %+v", tr.Content)
+	}
+}
+
+// TestToolResultsInHistoryUserMessage verifies that when there are multiple user messages
+// after tool results, the tool results are attached to the correct user message in history.
+func TestToolResultsInHistoryUserMessage(t *testing.T) {
+	// Sequence: user -> assistant (with tool_calls) -> tool (result) -> user -> assistant -> user
+	// The first user after tool should have tool results in history
+	input := []byte(`{
+		"model": "kiro-claude-opus-4-5-agentic",
+		"messages": [
+			{"role": "user", "content": "Hello"},
+			{
+				"role": "assistant",
+				"content": "I'll read the file.",
+				"tool_calls": [
+					{
+						"id": "call_1",
+						"type": "function",
+						"function": {
+							"name": "Read",
+							"arguments": "{}"
+						}
+					}
+				]
+			},
+			{
+				"role": "tool",
+				"tool_call_id": "call_1",
+				"content": "File result"
+			},
+			{"role": "user", "content": "Thanks for the file"},
+			{"role": "assistant", "content": "You're welcome"},
+			{"role": "user", "content": "Bye"}
+		]
+	}`)
+
+	result, _ := BuildKiroPayloadFromOpenAI(input, "kiro-model", "", "CLI", false, false, nil, nil)
+
+	var payload KiroPayload
+	if err := json.Unmarshal(result, &payload); err != nil {
+		t.Fatalf("Failed to unmarshal result: %v", err)
+	}
+
+	// History should have: user, assistant, user (with tool results), assistant
+	// CurrentMessage should be: last user "Bye"
+	t.Logf("History count: %d", len(payload.ConversationState.History))
+
+	// Find the user message in history with tool results
+	foundToolResults := false
+	for i, h := range payload.ConversationState.History {
+		if h.UserInputMessage != nil {
+			t.Logf("History[%d]: user message content=%q", i, h.UserInputMessage.Content)
+			if h.UserInputMessage.UserInputMessageContext != nil {
+				if len(h.UserInputMessage.UserInputMessageContext.ToolResults) > 0 {
+					foundToolResults = true
+					t.Logf("  Found %d tool results", len(h.UserInputMessage.UserInputMessageContext.ToolResults))
+					tr := h.UserInputMessage.UserInputMessageContext.ToolResults[0]
+					if tr.ToolUseID != "call_1" {
+						t.Errorf("Expected toolUseId 'call_1', got '%s'", tr.ToolUseID)
+					}
+				}
+			}
+		}
+		if h.AssistantResponseMessage != nil {
+			t.Logf("History[%d]: assistant message content=%q", i, h.AssistantResponseMessage.Content)
+		}
+	}
+
+	if !foundToolResults {
+		t.Error("Tool results were not attached to any user message in history")
+	}
+}
+
+// TestToolResultsWithMultipleToolCalls verifies handling of multiple tool calls
+func TestToolResultsWithMultipleToolCalls(t *testing.T) {
+	input := []byte(`{
+		"model": "kiro-claude-opus-4-5-agentic",
+		"messages": [
+			{"role": "user", "content": "Read two files for me"},
+			{
+				"role": "assistant",
+				"content": "I'll read both files.",
+				"tool_calls": [
+					{
+						"id": "call_1",
+						"type": "function",
+						"function": {
+							"name": "Read",
+							"arguments": "{\"file_path\": \"/tmp/file1.txt\"}"
+						}
+					},
+					{
+						"id": "call_2",
+						"type": "function",
+						"function": {
+							"name": "Read",
+							"arguments": "{\"file_path\": \"/tmp/file2.txt\"}"
+						}
+					}
+				]
+			},
+			{
+				"role": "tool",
+				"tool_call_id": "call_1",
+				"content": "Content of file 1"
+			},
+			{
+				"role": "tool",
+				"tool_call_id": "call_2",
+				"content": "Content of file 2"
+			},
+			{"role": "user", "content": "What do they say?"}
+		]
+	}`)
+
+	result, _ := BuildKiroPayloadFromOpenAI(input, "kiro-model", "", "CLI", false, false, nil, nil)
+
+	var payload KiroPayload
+	if err := json.Unmarshal(result, &payload); err != nil {
+		t.Fatalf("Failed to unmarshal result: %v", err)
+	}
+
+	t.Logf("History count: %d", len(payload.ConversationState.History))
+	t.Logf("CurrentMessage content: %q", payload.ConversationState.CurrentMessage.UserInputMessage.Content)
+
+	// Check if there are any tool results anywhere
+	var totalToolResults int
+	for i, h := range payload.ConversationState.History {
+		if h.UserInputMessage != nil && h.UserInputMessage.UserInputMessageContext != nil {
+			count := len(h.UserInputMessage.UserInputMessageContext.ToolResults)
+			t.Logf("History[%d] user message has %d tool results", i, count)
+			totalToolResults += count
+		}
+	}
+
+	ctx := payload.ConversationState.CurrentMessage.UserInputMessage.UserInputMessageContext
+	if ctx != nil {
+		t.Logf("CurrentMessage has %d tool results", len(ctx.ToolResults))
+		totalToolResults += len(ctx.ToolResults)
+	} else {
+		t.Logf("CurrentMessage has no UserInputMessageContext")
+	}
+
+	if totalToolResults != 2 {
+		t.Errorf("Expected 2 tool results total, got %d", totalToolResults)
+	}
+}
+
+// TestToolResultsAtEndOfConversation verifies tool results are handled when
+// the conversation ends with tool results (no following user message)
+func TestToolResultsAtEndOfConversation(t *testing.T) {
+	input := []byte(`{
+		"model": "kiro-claude-opus-4-5-agentic",
+		"messages": [
+			{"role": "user", "content": "Read a file"},
+			{
+				"role": "assistant",
+				"content": "Reading the file.",
+				"tool_calls": [
+					{
+						"id": "call_end",
+						"type": "function",
+						"function": {
+							"name": "Read",
+							"arguments": "{\"file_path\": \"/tmp/test.txt\"}"
+						}
+					}
+				]
+			},
+			{
+				"role": "tool",
+				"tool_call_id": "call_end",
+				"content": "File contents here"
+			}
+		]
+	}`)
+
+	result, _ := BuildKiroPayloadFromOpenAI(input, "kiro-model", "", "CLI", false, false, nil, nil)
+
+	var payload KiroPayload
+	if err := json.Unmarshal(result, &payload); err != nil {
+		t.Fatalf("Failed to unmarshal result: %v", err)
+	}
+
+	// When the last message is a tool result, a synthetic user message is created
+	// and tool results should be attached to it
+	ctx := payload.ConversationState.CurrentMessage.UserInputMessage.UserInputMessageContext
+	if ctx == nil || len(ctx.ToolResults) == 0 {
+		t.Error("Expected tool results to be attached to current message when conversation ends with tool result")
+	} else {
+		if ctx.ToolResults[0].ToolUseID != "call_end" {
+			t.Errorf("Expected toolUseId 'call_end', got '%s'", ctx.ToolResults[0].ToolUseID)
+		}
+	}
+}
+
+// TestToolResultsFollowedByAssistant verifies handling when tool results are followed
+// by an assistant message (no intermediate user message).
+// This is the pattern from LiteLLM translation of Anthropic format where:
+// user message has ONLY tool_result blocks -> LiteLLM creates tool messages
+// then the next message is assistant
+func TestToolResultsFollowedByAssistant(t *testing.T) {
+	// Sequence: user -> assistant (with tool_calls) -> tool -> tool -> assistant -> user
+	// This simulates LiteLLM's translation of:
+	//   user: "Read files"
+	//   assistant: [tool_use, tool_use]
+	//   user: [tool_result, tool_result]  <- becomes multiple "tool" role messages
+	//   assistant: "I've read them"
+	//   user: "What did they say?"
+	input := []byte(`{
+		"model": "kiro-claude-opus-4-5-agentic",
+		"messages": [
+			{"role": "user", "content": "Read two files for me"},
+			{
+				"role": "assistant",
+				"content": "I'll read both files.",
+				"tool_calls": [
+					{
+						"id": "call_1",
+						"type": "function",
+						"function": {
+							"name": "Read",
+							"arguments": "{\"file_path\": \"/tmp/a.txt\"}"
+						}
+					},
+					{
+						"id": "call_2",
+						"type": "function",
+						"function": {
+							"name": "Read",
+							"arguments": "{\"file_path\": \"/tmp/b.txt\"}"
+						}
+					}
+				]
+			},
+			{
+				"role": "tool",
+				"tool_call_id": "call_1",
+				"content": "Contents of file A"
+			},
+			{
+				"role": "tool",
+				"tool_call_id": "call_2",
+				"content": "Contents of file B"
+			},
+			{
+				"role": "assistant",
+				"content": "I've read both files."
+			},
+			{"role": "user", "content": "What did they say?"}
+		]
+	}`)
+
+	result, _ := BuildKiroPayloadFromOpenAI(input, "kiro-model", "", "CLI", false, false, nil, nil)
+
+	var payload KiroPayload
+	if err := json.Unmarshal(result, &payload); err != nil {
+		t.Fatalf("Failed to unmarshal result: %v", err)
+	}
+
+	t.Logf("History count: %d", len(payload.ConversationState.History))
+
+	// Tool results should be attached to a synthetic user message or the history should be valid
+	var totalToolResults int
+	for i, h := range payload.ConversationState.History {
+		if h.UserInputMessage != nil {
+			t.Logf("History[%d]: user message content=%q", i, h.UserInputMessage.Content)
+			if h.UserInputMessage.UserInputMessageContext != nil {
+				count := len(h.UserInputMessage.UserInputMessageContext.ToolResults)
+				t.Logf("  Has %d tool results", count)
+				totalToolResults += count
+			}
+		}
+		if h.AssistantResponseMessage != nil {
+			t.Logf("History[%d]: assistant message content=%q", i, h.AssistantResponseMessage.Content)
+		}
+	}
+
+	ctx := payload.ConversationState.CurrentMessage.UserInputMessage.UserInputMessageContext
+	if ctx != nil {
+		t.Logf("CurrentMessage has %d tool results", len(ctx.ToolResults))
+		totalToolResults += len(ctx.ToolResults)
+	}
+
+	if totalToolResults != 2 {
+		t.Errorf("Expected 2 tool results total, got %d", totalToolResults)
+	}
+}
+
+// TestAssistantEndsConversation verifies handling when assistant is the last message
+func TestAssistantEndsConversation(t *testing.T) {
+	input := []byte(`{
+		"model": "kiro-claude-opus-4-5-agentic",
+		"messages": [
+			{"role": "user", "content": "Hello"},
+			{
+				"role": "assistant",
+				"content": "Hi there!"
+			}
+		]
+	}`)
+
+	result, _ := BuildKiroPayloadFromOpenAI(input, "kiro-model", "", "CLI", false, false, nil, nil)
+
+	var payload KiroPayload
+	if err := json.Unmarshal(result, &payload); err != nil {
+		t.Fatalf("Failed to unmarshal result: %v", err)
+	}
+
+	// When assistant is last, a "Continue" user message should be created
+	if payload.ConversationState.CurrentMessage.UserInputMessage.Content == "" {
+		t.Error("Expected a 'Continue' message to be created when assistant is last")
+	}
+}

internal/translator/openai/claude/openai_claude_request.go

@@ -7,7 +7,6 @@ package claude
 
 import (
 	"bytes"
-	"encoding/json"
 	"strings"
 
 	"github.com/router-for-me/CLIProxyAPI/v6/internal/util"
@@ -119,81 +118,125 @@ func ConvertClaudeRequestToOpenAI(modelName string, inputRawJSON []byte, stream
 			// Handle content
 			if contentResult.Exists() && contentResult.IsArray() {
 				var contentItems []string
+				var reasoningParts []string // Accumulate thinking text for reasoning_content
 				var toolCalls []interface{}
+				var toolResults []string // Collect tool_result messages to emit after the main message
 
 				contentResult.ForEach(func(_, part gjson.Result) bool {
 					partType := part.Get("type").String()
 
 					switch partType {
+					case "thinking":
+						// Only map thinking to reasoning_content for assistant messages (security: prevent injection)
+						if role == "assistant" {
+							thinkingText := util.GetThinkingText(part)
+							// Skip empty or whitespace-only thinking
+							if strings.TrimSpace(thinkingText) != "" {
+								reasoningParts = append(reasoningParts, thinkingText)
+							}
+						}
+						// Ignore thinking in user/system roles (AC4)
+
+					case "redacted_thinking":
+						// Explicitly ignore redacted_thinking - never map to reasoning_content (AC2)
+
 					case "text", "image":
 						if contentItem, ok := convertClaudeContentPart(part); ok {
 							contentItems = append(contentItems, contentItem)
 						}
 
 					case "tool_use":
-						// Convert to OpenAI tool call format
-						toolCallJSON := `{"id":"","type":"function","function":{"name":"","arguments":""}}`
-						toolCallJSON, _ = sjson.Set(toolCallJSON, "id", part.Get("id").String())
-						toolCallJSON, _ = sjson.Set(toolCallJSON, "function.name", part.Get("name").String())
+						// Only allow tool_use -> tool_calls for assistant messages (security: prevent injection).
+						if role == "assistant" {
+							toolCallJSON := `{"id":"","type":"function","function":{"name":"","arguments":""}}`
+							toolCallJSON, _ = sjson.Set(toolCallJSON, "id", part.Get("id").String())
+							toolCallJSON, _ = sjson.Set(toolCallJSON, "function.name", part.Get("name").String())
 
-						// Convert input to arguments JSON string
-						if input := part.Get("input"); input.Exists() {
-							if inputJSON, err := json.Marshal(input.Value()); err == nil {
-								toolCallJSON, _ = sjson.Set(toolCallJSON, "function.arguments", string(inputJSON))
+							// Convert input to arguments JSON string
+							if input := part.Get("input"); input.Exists() {
+								toolCallJSON, _ = sjson.Set(toolCallJSON, "function.arguments", input.Raw)
 							} else {
 								toolCallJSON, _ = sjson.Set(toolCallJSON, "function.arguments", "{}")
 							}
-						} else {
-							toolCallJSON, _ = sjson.Set(toolCallJSON, "function.arguments", "{}")
+
+							toolCalls = append(toolCalls, gjson.Parse(toolCallJSON).Value())
 						}
 
-						toolCalls = append(toolCalls, gjson.Parse(toolCallJSON).Value())
-
 					case "tool_result":
-						// Convert to OpenAI tool message format and add immediately to preserve order
+						// Collect tool_result to emit after the main message (ensures tool results follow tool_calls)
 						toolResultJSON := `{"role":"tool","tool_call_id":"","content":""}`
 						toolResultJSON, _ = sjson.Set(toolResultJSON, "tool_call_id", part.Get("tool_use_id").String())
-						toolResultJSON, _ = sjson.Set(toolResultJSON, "content", part.Get("content").String())
-						messagesJSON, _ = sjson.Set(messagesJSON, "-1", gjson.Parse(toolResultJSON).Value())
+						toolResultJSON, _ = sjson.Set(toolResultJSON, "content", convertClaudeToolResultContentToString(part.Get("content")))
+						toolResults = append(toolResults, toolResultJSON)
 					}
 					return true
 				})
 
-				// Emit text/image content as one message
-				if len(contentItems) > 0 {
-					msgJSON := `{"role":"","content":""}`
-					msgJSON, _ = sjson.Set(msgJSON, "role", role)
-
-					contentArrayJSON := "[]"
-					for _, contentItem := range contentItems {
-						contentArrayJSON, _ = sjson.SetRaw(contentArrayJSON, "-1", contentItem)
-					}
-					msgJSON, _ = sjson.SetRaw(msgJSON, "content", contentArrayJSON)
-
-					contentValue := gjson.Get(msgJSON, "content")
-					hasContent := false
-					switch {
-					case !contentValue.Exists():
-						hasContent = false
-					case contentValue.Type == gjson.String:
-						hasContent = contentValue.String() != ""
-					case contentValue.IsArray():
-						hasContent = len(contentValue.Array()) > 0
-					default:
-						hasContent = contentValue.Raw != "" && contentValue.Raw != "null"
-					}
-
-					if hasContent {
-						messagesJSON, _ = sjson.Set(messagesJSON, "-1", gjson.Parse(msgJSON).Value())
-					}
+				// Build reasoning content string
+				reasoningContent := ""
+				if len(reasoningParts) > 0 {
+					reasoningContent = strings.Join(reasoningParts, "\n\n")
 				}
 
-				// Emit tool calls in a separate assistant message
-				if role == "assistant" && len(toolCalls) > 0 {
-					toolCallMsgJSON := `{"role":"assistant","tool_calls":[]}`
-					toolCallsJSON, _ := json.Marshal(toolCalls)
-					toolCallMsgJSON, _ = sjson.SetRaw(toolCallMsgJSON, "tool_calls", string(toolCallsJSON))
-					messagesJSON, _ = sjson.Set(messagesJSON, "-1", gjson.Parse(toolCallMsgJSON).Value())
+				hasContent := len(contentItems) > 0
+				hasReasoning := reasoningContent != ""
+				hasToolCalls := len(toolCalls) > 0
+				hasToolResults := len(toolResults) > 0
+
+				// OpenAI requires: tool messages MUST immediately follow the assistant message with tool_calls.
+				// Therefore, we emit tool_result messages FIRST (they respond to the previous assistant's tool_calls),
+				// then emit the current message's content.
+				for _, toolResultJSON := range toolResults {
+					messagesJSON, _ = sjson.Set(messagesJSON, "-1", gjson.Parse(toolResultJSON).Value())
+				}
+
+				// For assistant messages: emit a single unified message with content, tool_calls, and reasoning_content
+				// This avoids splitting into multiple assistant messages which breaks OpenAI tool-call adjacency
+				if role == "assistant" {
+					if hasContent || hasReasoning || hasToolCalls {
+						msgJSON := `{"role":"assistant"}`
+
+						// Add content (as array if we have items, empty string if reasoning-only)
+						if hasContent {
+							contentArrayJSON := "[]"
+							for _, contentItem := range contentItems {
+								contentArrayJSON, _ = sjson.SetRaw(contentArrayJSON, "-1", contentItem)
+							}
+							msgJSON, _ = sjson.SetRaw(msgJSON, "content", contentArrayJSON)
+						} else {
+							// Ensure content field exists for OpenAI compatibility
+							msgJSON, _ = sjson.Set(msgJSON, "content", "")
+						}
+
+						// Add reasoning_content if present
+						if hasReasoning {
+							msgJSON, _ = sjson.Set(msgJSON, "reasoning_content", reasoningContent)
+						}
+
+						// Add tool_calls if present (in same message as content)
+						if hasToolCalls {
+							msgJSON, _ = sjson.Set(msgJSON, "tool_calls", toolCalls)
+						}
+
+						messagesJSON, _ = sjson.Set(messagesJSON, "-1", gjson.Parse(msgJSON).Value())
+					}
+				} else {
+					// For non-assistant roles: emit content message if we have content
+					// If the message only contains tool_results (no text/image), we still processed them above
+					if hasContent {
+						msgJSON := `{"role":""}`
+						msgJSON, _ = sjson.Set(msgJSON, "role", role)
+
+						contentArrayJSON := "[]"
+						for _, contentItem := range contentItems {
+							contentArrayJSON, _ = sjson.SetRaw(contentArrayJSON, "-1", contentItem)
+						}
+						msgJSON, _ = sjson.SetRaw(msgJSON, "content", contentArrayJSON)
+
+						messagesJSON, _ = sjson.Set(messagesJSON, "-1", gjson.Parse(msgJSON).Value())
+					} else if hasToolResults && !hasContent {
+						// tool_results already emitted above, no additional user message needed
+					}
 				}
 
 			} else if contentResult.Exists() && contentResult.Type == gjson.String {
@@ -313,3 +356,43 @@ func convertClaudeContentPart(part gjson.Result) (string, bool) {
 		return "", false
 	}
 }
+
+func convertClaudeToolResultContentToString(content gjson.Result) string {
+	if !content.Exists() {
+		return ""
+	}
+
+	if content.Type == gjson.String {
+		return content.String()
+	}
+
+	if content.IsArray() {
+		var parts []string
+		content.ForEach(func(_, item gjson.Result) bool {
+			switch {
+			case item.Type == gjson.String:
+				parts = append(parts, item.String())
+			case item.IsObject() && item.Get("text").Exists() && item.Get("text").Type == gjson.String:
+				parts = append(parts, item.Get("text").String())
+			default:
+				parts = append(parts, item.Raw)
+			}
+			return true
+		})
+
+		joined := strings.Join(parts, "\n\n")
+		if strings.TrimSpace(joined) != "" {
+			return joined
+		}
+		return content.Raw
+	}
+
+	if content.IsObject() {
+		if text := content.Get("text"); text.Exists() && text.Type == gjson.String {
+			return text.String()
+		}
+		return content.Raw
+	}
+
+	return content.Raw
+}

internal/translator/openai/claude/openai_claude_request_test.go

@@ -0,0 +1,500 @@
+package claude
+
+import (
+	"testing"
+
+	"github.com/tidwall/gjson"
+)
+
+// TestConvertClaudeRequestToOpenAI_ThinkingToReasoningContent tests the mapping
+// of Claude thinking content to OpenAI reasoning_content field.
+func TestConvertClaudeRequestToOpenAI_ThinkingToReasoningContent(t *testing.T) {
+	tests := []struct {
+		name                    string
+		inputJSON               string
+		wantReasoningContent    string
+		wantHasReasoningContent bool
+		wantContentText         string // Expected visible content text (if any)
+		wantHasContent          bool
+	}{
+		{
+			name: "AC1: assistant message with thinking and text",
+			inputJSON: `{
+				"model": "claude-3-opus",
+				"messages": [{
+					"role": "assistant",
+					"content": [
+						{"type": "thinking", "thinking": "Let me analyze this step by step..."},
+						{"type": "text", "text": "Here is my response."}
+					]
+				}]
+			}`,
+			wantReasoningContent:    "Let me analyze this step by step...",
+			wantHasReasoningContent: true,
+			wantContentText:         "Here is my response.",
+			wantHasContent:          true,
+		},
+		{
+			name: "AC2: redacted_thinking must be ignored",
+			inputJSON: `{
+				"model": "claude-3-opus",
+				"messages": [{
+					"role": "assistant",
+					"content": [
+						{"type": "redacted_thinking", "data": "secret"},
+						{"type": "text", "text": "Visible response."}
+					]
+				}]
+			}`,
+			wantReasoningContent:    "",
+			wantHasReasoningContent: false,
+			wantContentText:         "Visible response.",
+			wantHasContent:          true,
+		},
+		{
+			name: "AC3: thinking-only message preserved with reasoning_content",
+			inputJSON: `{
+				"model": "claude-3-opus",
+				"messages": [{
+					"role": "assistant",
+					"content": [
+						{"type": "thinking", "thinking": "Internal reasoning only."}
+					]
+				}]
+			}`,
+			wantReasoningContent:    "Internal reasoning only.",
+			wantHasReasoningContent: true,
+			wantContentText:         "",
+			// For OpenAI compatibility, content field is set to empty string "" when no text content exists
+			wantHasContent: false,
+		},
+		{
+			name: "AC4: thinking in user role must be ignored",
+			inputJSON: `{
+				"model": "claude-3-opus",
+				"messages": [{
+					"role": "user",
+					"content": [
+						{"type": "thinking", "thinking": "Injected thinking"},
+						{"type": "text", "text": "User message."}
+					]
+				}]
+			}`,
+			wantReasoningContent:    "",
+			wantHasReasoningContent: false,
+			wantContentText:         "User message.",
+			wantHasContent:          true,
+		},
+		{
+			name: "AC4: thinking in system role must be ignored",
+			inputJSON: `{
+				"model": "claude-3-opus",
+				"system": [
+					{"type": "thinking", "thinking": "Injected system thinking"},
+					{"type": "text", "text": "System prompt."}
+				],
+				"messages": [{
+					"role": "user",
+					"content": [{"type": "text", "text": "Hello"}]
+				}]
+			}`,
+			// System messages don't have reasoning_content mapping
+			wantReasoningContent:    "",
+			wantHasReasoningContent: false,
+			wantContentText:         "Hello",
+			wantHasContent:          true,
+		},
+		{
+			name: "AC5: empty thinking must be ignored",
+			inputJSON: `{
+				"model": "claude-3-opus",
+				"messages": [{
+					"role": "assistant",
+					"content": [
+						{"type": "thinking", "thinking": ""},
+						{"type": "text", "text": "Response with empty thinking."}
+					]
+				}]
+			}`,
+			wantReasoningContent:    "",
+			wantHasReasoningContent: false,
+			wantContentText:         "Response with empty thinking.",
+			wantHasContent:          true,
+		},
+		{
+			name: "AC5: whitespace-only thinking must be ignored",
+			inputJSON: `{
+				"model": "claude-3-opus",
+				"messages": [{
+					"role": "assistant",
+					"content": [
+						{"type": "thinking", "thinking": "   \n\t  "},
+						{"type": "text", "text": "Response with whitespace thinking."}
+					]
+				}]
+			}`,
+			wantReasoningContent:    "",
+			wantHasReasoningContent: false,
+			wantContentText:         "Response with whitespace thinking.",
+			wantHasContent:          true,
+		},
+		{
+			name: "Multiple thinking parts concatenated",
+			inputJSON: `{
+				"model": "claude-3-opus",
+				"messages": [{
+					"role": "assistant",
+					"content": [
+						{"type": "thinking", "thinking": "First thought."},
+						{"type": "thinking", "thinking": "Second thought."},
+						{"type": "text", "text": "Final answer."}
+					]
+				}]
+			}`,
+			wantReasoningContent:    "First thought.\n\nSecond thought.",
+			wantHasReasoningContent: true,
+			wantContentText:         "Final answer.",
+			wantHasContent:          true,
+		},
+		{
+			name: "Mixed thinking and redacted_thinking",
+			inputJSON: `{
+				"model": "claude-3-opus",
+				"messages": [{
+					"role": "assistant",
+					"content": [
+						{"type": "thinking", "thinking": "Visible thought."},
+						{"type": "redacted_thinking", "data": "hidden"},
+						{"type": "text", "text": "Answer."}
+					]
+				}]
+			}`,
+			wantReasoningContent:    "Visible thought.",
+			wantHasReasoningContent: true,
+			wantContentText:         "Answer.",
+			wantHasContent:          true,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			result := ConvertClaudeRequestToOpenAI("test-model", []byte(tt.inputJSON), false)
+			resultJSON := gjson.ParseBytes(result)
+
+			// Find the relevant message (skip system message at index 0)
+			messages := resultJSON.Get("messages").Array()
+			if len(messages) < 2 {
+				if tt.wantHasReasoningContent || tt.wantHasContent {
+					t.Fatalf("Expected at least 2 messages (system + user/assistant), got %d", len(messages))
+				}
+				return
+			}
+
+			// Check the last non-system message
+			var targetMsg gjson.Result
+			for i := len(messages) - 1; i >= 0; i-- {
+				if messages[i].Get("role").String() != "system" {
+					targetMsg = messages[i]
+					break
+				}
+			}
+
+			// Check reasoning_content
+			gotReasoningContent := targetMsg.Get("reasoning_content").String()
+			gotHasReasoningContent := targetMsg.Get("reasoning_content").Exists()
+
+			if gotHasReasoningContent != tt.wantHasReasoningContent {
+				t.Errorf("reasoning_content existence = %v, want %v", gotHasReasoningContent, tt.wantHasReasoningContent)
+			}
+
+			if gotReasoningContent != tt.wantReasoningContent {
+				t.Errorf("reasoning_content = %q, want %q", gotReasoningContent, tt.wantReasoningContent)
+			}
+
+			// Check content
+			content := targetMsg.Get("content")
+			// content has meaningful content if it's a non-empty array, or a non-empty string
+			var gotHasContent bool
+			switch {
+			case content.IsArray():
+				gotHasContent = len(content.Array()) > 0
+			case content.Type == gjson.String:
+				gotHasContent = content.String() != ""
+			default:
+				gotHasContent = false
+			}
+
+			if gotHasContent != tt.wantHasContent {
+				t.Errorf("content existence = %v, want %v", gotHasContent, tt.wantHasContent)
+			}
+
+			if tt.wantHasContent && tt.wantContentText != "" {
+				// Find text content
+				var foundText string
+				content.ForEach(func(_, v gjson.Result) bool {
+					if v.Get("type").String() == "text" {
+						foundText = v.Get("text").String()
+						return false
+					}
+					return true
+				})
+				if foundText != tt.wantContentText {
+					t.Errorf("content text = %q, want %q", foundText, tt.wantContentText)
+				}
+			}
+		})
+	}
+}
+
+// TestConvertClaudeRequestToOpenAI_ThinkingOnlyMessagePreserved tests AC3:
+// that a message with only thinking content is preserved (not dropped).
+func TestConvertClaudeRequestToOpenAI_ThinkingOnlyMessagePreserved(t *testing.T) {
+	inputJSON := `{
+		"model": "claude-3-opus",
+		"messages": [
+			{
+				"role": "user",
+				"content": [{"type": "text", "text": "What is 2+2?"}]
+			},
+			{
+				"role": "assistant",
+				"content": [{"type": "thinking", "thinking": "Let me calculate: 2+2=4"}]
+			},
+			{
+				"role": "user",
+				"content": [{"type": "text", "text": "Thanks"}]
+			}
+		]
+	}`
+
+	result := ConvertClaudeRequestToOpenAI("test-model", []byte(inputJSON), false)
+	resultJSON := gjson.ParseBytes(result)
+
+	messages := resultJSON.Get("messages").Array()
+
+	// Should have: system (auto-added) + user + assistant (thinking-only) + user = 4 messages
+	if len(messages) != 4 {
+		t.Fatalf("Expected 4 messages, got %d. Messages: %v", len(messages), resultJSON.Get("messages").Raw)
+	}
+
+	// Check the assistant message (index 2) has reasoning_content
+	assistantMsg := messages[2]
+	if assistantMsg.Get("role").String() != "assistant" {
+		t.Errorf("Expected message[2] to be assistant, got %s", assistantMsg.Get("role").String())
+	}
+
+	if !assistantMsg.Get("reasoning_content").Exists() {
+		t.Error("Expected assistant message to have reasoning_content")
+	}
+
+	if assistantMsg.Get("reasoning_content").String() != "Let me calculate: 2+2=4" {
+		t.Errorf("Unexpected reasoning_content: %s", assistantMsg.Get("reasoning_content").String())
+	}
+}
+
+func TestConvertClaudeRequestToOpenAI_ToolResultOrderAndContent(t *testing.T) {
+	inputJSON := `{
+		"model": "claude-3-opus",
+		"messages": [
+			{
+				"role": "assistant",
+				"content": [
+					{"type": "tool_use", "id": "call_1", "name": "do_work", "input": {"a": 1}}
+				]
+			},
+			{
+				"role": "user",
+				"content": [
+					{"type": "text", "text": "before"},
+					{"type": "tool_result", "tool_use_id": "call_1", "content": [{"type":"text","text":"tool ok"}]},
+					{"type": "text", "text": "after"}
+				]
+			}
+		]
+	}`
+
+	result := ConvertClaudeRequestToOpenAI("test-model", []byte(inputJSON), false)
+	resultJSON := gjson.ParseBytes(result)
+	messages := resultJSON.Get("messages").Array()
+
+	// OpenAI requires: tool messages MUST immediately follow assistant(tool_calls).
+	// Correct order: system + assistant(tool_calls) + tool(result) + user(before+after)
+	if len(messages) != 4 {
+		t.Fatalf("Expected 4 messages, got %d. Messages: %s", len(messages), resultJSON.Get("messages").Raw)
+	}
+
+	if messages[0].Get("role").String() != "system" {
+		t.Fatalf("Expected messages[0] to be system, got %s", messages[0].Get("role").String())
+	}
+
+	if messages[1].Get("role").String() != "assistant" || !messages[1].Get("tool_calls").Exists() {
+		t.Fatalf("Expected messages[1] to be assistant tool_calls, got %s: %s", messages[1].Get("role").String(), messages[1].Raw)
+	}
+
+	// tool message MUST immediately follow assistant(tool_calls) per OpenAI spec
+	if messages[2].Get("role").String() != "tool" {
+		t.Fatalf("Expected messages[2] to be tool (must follow tool_calls), got %s", messages[2].Get("role").String())
+	}
+	if got := messages[2].Get("tool_call_id").String(); got != "call_1" {
+		t.Fatalf("Expected tool_call_id %q, got %q", "call_1", got)
+	}
+	if got := messages[2].Get("content").String(); got != "tool ok" {
+		t.Fatalf("Expected tool content %q, got %q", "tool ok", got)
+	}
+
+	// User message comes after tool message
+	if messages[3].Get("role").String() != "user" {
+		t.Fatalf("Expected messages[3] to be user, got %s", messages[3].Get("role").String())
+	}
+	// User message should contain both "before" and "after" text
+	if got := messages[3].Get("content.0.text").String(); got != "before" {
+		t.Fatalf("Expected user text[0] %q, got %q", "before", got)
+	}
+	if got := messages[3].Get("content.1.text").String(); got != "after" {
+		t.Fatalf("Expected user text[1] %q, got %q", "after", got)
+	}
+}
+
+func TestConvertClaudeRequestToOpenAI_ToolResultObjectContent(t *testing.T) {
+	inputJSON := `{
+		"model": "claude-3-opus",
+		"messages": [
+			{
+				"role": "assistant",
+				"content": [
+					{"type": "tool_use", "id": "call_1", "name": "do_work", "input": {"a": 1}}
+				]
+			},
+			{
+				"role": "user",
+				"content": [
+					{"type": "tool_result", "tool_use_id": "call_1", "content": {"foo": "bar"}}
+				]
+			}
+		]
+	}`
+
+	result := ConvertClaudeRequestToOpenAI("test-model", []byte(inputJSON), false)
+	resultJSON := gjson.ParseBytes(result)
+	messages := resultJSON.Get("messages").Array()
+
+	// system + assistant(tool_calls) + tool(result)
+	if len(messages) != 3 {
+		t.Fatalf("Expected 3 messages, got %d. Messages: %s", len(messages), resultJSON.Get("messages").Raw)
+	}
+
+	if messages[2].Get("role").String() != "tool" {
+		t.Fatalf("Expected messages[2] to be tool, got %s", messages[2].Get("role").String())
+	}
+
+	toolContent := messages[2].Get("content").String()
+	parsed := gjson.Parse(toolContent)
+	if parsed.Get("foo").String() != "bar" {
+		t.Fatalf("Expected tool content JSON foo=bar, got %q", toolContent)
+	}
+}
+
+func TestConvertClaudeRequestToOpenAI_AssistantTextToolUseTextOrder(t *testing.T) {
+	inputJSON := `{
+		"model": "claude-3-opus",
+		"messages": [
+			{
+				"role": "assistant",
+				"content": [
+					{"type": "text", "text": "pre"},
+					{"type": "tool_use", "id": "call_1", "name": "do_work", "input": {"a": 1}},
+					{"type": "text", "text": "post"}
+				]
+			}
+		]
+	}`
+
+	result := ConvertClaudeRequestToOpenAI("test-model", []byte(inputJSON), false)
+	resultJSON := gjson.ParseBytes(result)
+	messages := resultJSON.Get("messages").Array()
+
+	// New behavior: content + tool_calls unified in single assistant message
+	// Expect: system + assistant(content[pre,post] + tool_calls)
+	if len(messages) != 2 {
+		t.Fatalf("Expected 2 messages, got %d. Messages: %s", len(messages), resultJSON.Get("messages").Raw)
+	}
+
+	if messages[0].Get("role").String() != "system" {
+		t.Fatalf("Expected messages[0] to be system, got %s", messages[0].Get("role").String())
+	}
+
+	assistantMsg := messages[1]
+	if assistantMsg.Get("role").String() != "assistant" {
+		t.Fatalf("Expected messages[1] to be assistant, got %s", assistantMsg.Get("role").String())
+	}
+
+	// Should have both content and tool_calls in same message
+	if !assistantMsg.Get("tool_calls").Exists() {
+		t.Fatalf("Expected assistant message to have tool_calls")
+	}
+	if got := assistantMsg.Get("tool_calls.0.id").String(); got != "call_1" {
+		t.Fatalf("Expected tool_call id %q, got %q", "call_1", got)
+	}
+	if got := assistantMsg.Get("tool_calls.0.function.name").String(); got != "do_work" {
+		t.Fatalf("Expected tool_call name %q, got %q", "do_work", got)
+	}
+
+	// Content should have both pre and post text
+	if got := assistantMsg.Get("content.0.text").String(); got != "pre" {
+		t.Fatalf("Expected content[0] text %q, got %q", "pre", got)
+	}
+	if got := assistantMsg.Get("content.1.text").String(); got != "post" {
+		t.Fatalf("Expected content[1] text %q, got %q", "post", got)
+	}
+}
+
+func TestConvertClaudeRequestToOpenAI_AssistantThinkingToolUseThinkingSplit(t *testing.T) {
+	inputJSON := `{
+		"model": "claude-3-opus",
+		"messages": [
+			{
+				"role": "assistant",
+				"content": [
+					{"type": "thinking", "thinking": "t1"},
+					{"type": "text", "text": "pre"},
+					{"type": "tool_use", "id": "call_1", "name": "do_work", "input": {"a": 1}},
+					{"type": "thinking", "thinking": "t2"},
+					{"type": "text", "text": "post"}
+				]
+			}
+		]
+	}`
+
+	result := ConvertClaudeRequestToOpenAI("test-model", []byte(inputJSON), false)
+	resultJSON := gjson.ParseBytes(result)
+	messages := resultJSON.Get("messages").Array()
+
+	// New behavior: all content, thinking, and tool_calls unified in single assistant message
+	// Expect: system + assistant(content[pre,post] + tool_calls + reasoning_content[t1+t2])
+	if len(messages) != 2 {
+		t.Fatalf("Expected 2 messages, got %d. Messages: %s", len(messages), resultJSON.Get("messages").Raw)
+	}
+
+	assistantMsg := messages[1]
+	if assistantMsg.Get("role").String() != "assistant" {
+		t.Fatalf("Expected messages[1] to be assistant, got %s", assistantMsg.Get("role").String())
+	}
+
+	// Should have content with both pre and post
+	if got := assistantMsg.Get("content.0.text").String(); got != "pre" {
+		t.Fatalf("Expected content[0] text %q, got %q", "pre", got)
+	}
+	if got := assistantMsg.Get("content.1.text").String(); got != "post" {
+		t.Fatalf("Expected content[1] text %q, got %q", "post", got)
+	}
+
+	// Should have tool_calls
+	if !assistantMsg.Get("tool_calls").Exists() {
+		t.Fatalf("Expected assistant message to have tool_calls")
+	}
+
+	// Should have combined reasoning_content from both thinking blocks
+	if got := assistantMsg.Get("reasoning_content").String(); got != "t1\n\nt2" {
+		t.Fatalf("Expected reasoning_content %q, got %q", "t1\n\nt2", got)
+	}
+}

internal/translator/openai/claude/openai_claude_response.go

@@ -8,7 +8,6 @@ package claude
 import (
 	"bytes"
 	"context"
-	"encoding/json"
 	"fmt"
 	"strings"
 
@@ -133,24 +132,10 @@ func convertOpenAIStreamingChunkToAnthropic(rawJSON []byte, param *ConvertOpenAI
 	if delta := root.Get("choices.0.delta"); delta.Exists() {
 		if !param.MessageStarted {
 			// Send message_start event
-			messageStart := map[string]interface{}{
-				"type": "message_start",
-				"message": map[string]interface{}{
-					"id":            param.MessageID,
-					"type":          "message",
-					"role":          "assistant",
-					"model":         param.Model,
-					"content":       []interface{}{},
-					"stop_reason":   nil,
-					"stop_sequence": nil,
-					"usage": map[string]interface{}{
-						"input_tokens":  0,
-						"output_tokens": 0,
-					},
-				},
-			}
-			messageStartJSON, _ := json.Marshal(messageStart)
-			results = append(results, "event: message_start\ndata: "+string(messageStartJSON)+"\n\n")
+			messageStartJSON := `{"type":"message_start","message":{"id":"","type":"message","role":"assistant","model":"","content":[],"stop_reason":null,"stop_sequence":null,"usage":{"input_tokens":0,"output_tokens":0}}}`
+			messageStartJSON, _ = sjson.Set(messageStartJSON, "message.id", param.MessageID)
+			messageStartJSON, _ = sjson.Set(messageStartJSON, "message.model", param.Model)
+			results = append(results, "event: message_start\ndata: "+messageStartJSON+"\n\n")
 			param.MessageStarted = true
 
 			// Don't send content_block_start for text here - wait for actual content
@@ -168,29 +153,16 @@ func convertOpenAIStreamingChunkToAnthropic(rawJSON []byte, param *ConvertOpenAI
 						param.ThinkingContentBlockIndex = param.NextContentBlockIndex
 						param.NextContentBlockIndex++
 					}
-					contentBlockStart := map[string]interface{}{
-						"type":  "content_block_start",
-						"index": param.ThinkingContentBlockIndex,
-						"content_block": map[string]interface{}{
-							"type":     "thinking",
-							"thinking": "",
-						},
-					}
-					contentBlockStartJSON, _ := json.Marshal(contentBlockStart)
-					results = append(results, "event: content_block_start\ndata: "+string(contentBlockStartJSON)+"\n\n")
+					contentBlockStartJSON := `{"type":"content_block_start","index":0,"content_block":{"type":"thinking","thinking":""}}`
+					contentBlockStartJSON, _ = sjson.Set(contentBlockStartJSON, "index", param.ThinkingContentBlockIndex)
+					results = append(results, "event: content_block_start\ndata: "+contentBlockStartJSON+"\n\n")
 					param.ThinkingContentBlockStarted = true
 				}
 
-				thinkingDelta := map[string]interface{}{
-					"type":  "content_block_delta",
-					"index": param.ThinkingContentBlockIndex,
-					"delta": map[string]interface{}{
-						"type":     "thinking_delta",
-						"thinking": reasoningText,
-					},
-				}
-				thinkingDeltaJSON, _ := json.Marshal(thinkingDelta)
-				results = append(results, "event: content_block_delta\ndata: "+string(thinkingDeltaJSON)+"\n\n")
+				thinkingDeltaJSON := `{"type":"content_block_delta","index":0,"delta":{"type":"thinking_delta","thinking":""}}`
+				thinkingDeltaJSON, _ = sjson.Set(thinkingDeltaJSON, "index", param.ThinkingContentBlockIndex)
+				thinkingDeltaJSON, _ = sjson.Set(thinkingDeltaJSON, "delta.thinking", reasoningText)
+				results = append(results, "event: content_block_delta\ndata: "+thinkingDeltaJSON+"\n\n")
 			}
 		}
 
@@ -203,29 +175,16 @@ func convertOpenAIStreamingChunkToAnthropic(rawJSON []byte, param *ConvertOpenAI
 					param.TextContentBlockIndex = param.NextContentBlockIndex
 					param.NextContentBlockIndex++
 				}
-				contentBlockStart := map[string]interface{}{
-					"type":  "content_block_start",
-					"index": param.TextContentBlockIndex,
-					"content_block": map[string]interface{}{
-						"type": "text",
-						"text": "",
-					},
-				}
-				contentBlockStartJSON, _ := json.Marshal(contentBlockStart)
-				results = append(results, "event: content_block_start\ndata: "+string(contentBlockStartJSON)+"\n\n")
+				contentBlockStartJSON := `{"type":"content_block_start","index":0,"content_block":{"type":"text","text":""}}`
+				contentBlockStartJSON, _ = sjson.Set(contentBlockStartJSON, "index", param.TextContentBlockIndex)
+				results = append(results, "event: content_block_start\ndata: "+contentBlockStartJSON+"\n\n")
 				param.TextContentBlockStarted = true
 			}
 
-			contentDelta := map[string]interface{}{
-				"type":  "content_block_delta",
-				"index": param.TextContentBlockIndex,
-				"delta": map[string]interface{}{
-					"type": "text_delta",
-					"text": content.String(),
-				},
-			}
-			contentDeltaJSON, _ := json.Marshal(contentDelta)
-			results = append(results, "event: content_block_delta\ndata: "+string(contentDeltaJSON)+"\n\n")
+			contentDeltaJSON := `{"type":"content_block_delta","index":0,"delta":{"type":"text_delta","text":""}}`
+			contentDeltaJSON, _ = sjson.Set(contentDeltaJSON, "index", param.TextContentBlockIndex)
+			contentDeltaJSON, _ = sjson.Set(contentDeltaJSON, "delta.text", content.String())
+			results = append(results, "event: content_block_delta\ndata: "+contentDeltaJSON+"\n\n")
 
 			// Accumulate content
 			param.ContentAccumulator.WriteString(content.String())
@@ -263,18 +222,11 @@ func convertOpenAIStreamingChunkToAnthropic(rawJSON []byte, param *ConvertOpenAI
 						stopTextContentBlock(param, &results)
 
 						// Send content_block_start for tool_use
-						contentBlockStart := map[string]interface{}{
-							"type":  "content_block_start",
-							"index": blockIndex,
-							"content_block": map[string]interface{}{
-								"type":  "tool_use",
-								"id":    accumulator.ID,
-								"name":  accumulator.Name,
-								"input": map[string]interface{}{},
-							},
-						}
-						contentBlockStartJSON, _ := json.Marshal(contentBlockStart)
-						results = append(results, "event: content_block_start\ndata: "+string(contentBlockStartJSON)+"\n\n")
+						contentBlockStartJSON := `{"type":"content_block_start","index":0,"content_block":{"type":"tool_use","id":"","name":"","input":{}}}`
+						contentBlockStartJSON, _ = sjson.Set(contentBlockStartJSON, "index", blockIndex)
+						contentBlockStartJSON, _ = sjson.Set(contentBlockStartJSON, "content_block.id", accumulator.ID)
+						contentBlockStartJSON, _ = sjson.Set(contentBlockStartJSON, "content_block.name", accumulator.Name)
+						results = append(results, "event: content_block_start\ndata: "+contentBlockStartJSON+"\n\n")
 					}
 
 					// Handle function arguments
@@ -298,12 +250,9 @@ func convertOpenAIStreamingChunkToAnthropic(rawJSON []byte, param *ConvertOpenAI
 
 		// Send content_block_stop for thinking content if needed
 		if param.ThinkingContentBlockStarted {
-			contentBlockStop := map[string]interface{}{
-				"type":  "content_block_stop",
-				"index": param.ThinkingContentBlockIndex,
-			}
-			contentBlockStopJSON, _ := json.Marshal(contentBlockStop)
-			results = append(results, "event: content_block_stop\ndata: "+string(contentBlockStopJSON)+"\n\n")
+			contentBlockStopJSON := `{"type":"content_block_stop","index":0}`
+			contentBlockStopJSON, _ = sjson.Set(contentBlockStopJSON, "index", param.ThinkingContentBlockIndex)
+			results = append(results, "event: content_block_stop\ndata: "+contentBlockStopJSON+"\n\n")
 			param.ThinkingContentBlockStarted = false
 			param.ThinkingContentBlockIndex = -1
 		}
@@ -319,24 +268,15 @@ func convertOpenAIStreamingChunkToAnthropic(rawJSON []byte, param *ConvertOpenAI
 
 				// Send complete input_json_delta with all accumulated arguments
 				if accumulator.Arguments.Len() > 0 {
-					inputDelta := map[string]interface{}{
-						"type":  "content_block_delta",
-						"index": blockIndex,
-						"delta": map[string]interface{}{
-							"type":         "input_json_delta",
-							"partial_json": util.FixJSON(accumulator.Arguments.String()),
-						},
-					}
-					inputDeltaJSON, _ := json.Marshal(inputDelta)
-					results = append(results, "event: content_block_delta\ndata: "+string(inputDeltaJSON)+"\n\n")
+					inputDeltaJSON := `{"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""}}`
+					inputDeltaJSON, _ = sjson.Set(inputDeltaJSON, "index", blockIndex)
+					inputDeltaJSON, _ = sjson.Set(inputDeltaJSON, "delta.partial_json", util.FixJSON(accumulator.Arguments.String()))
+					results = append(results, "event: content_block_delta\ndata: "+inputDeltaJSON+"\n\n")
 				}
 
-				contentBlockStop := map[string]interface{}{
-					"type":  "content_block_stop",
-					"index": blockIndex,
-				}
-				contentBlockStopJSON, _ := json.Marshal(contentBlockStop)
-				results = append(results, "event: content_block_stop\ndata: "+string(contentBlockStopJSON)+"\n\n")
+				contentBlockStopJSON := `{"type":"content_block_stop","index":0}`
+				contentBlockStopJSON, _ = sjson.Set(contentBlockStopJSON, "index", blockIndex)
+				results = append(results, "event: content_block_stop\ndata: "+contentBlockStopJSON+"\n\n")
 				delete(param.ToolCallBlockIndexes, index)
 			}
 			param.ContentBlocksStopped = true
@@ -361,20 +301,11 @@ func convertOpenAIStreamingChunkToAnthropic(rawJSON []byte, param *ConvertOpenAI
 			}
 		}
 		// Send message_delta with usage
-		messageDelta := map[string]interface{}{
-			"type": "message_delta",
-			"delta": map[string]interface{}{
-				"stop_reason":   mapOpenAIFinishReasonToAnthropic(param.FinishReason),
-				"stop_sequence": nil,
-			},
-			"usage": map[string]interface{}{
-				"input_tokens":  inputTokens,
-				"output_tokens": outputTokens,
-			},
-		}
-
-		messageDeltaJSON, _ := json.Marshal(messageDelta)
-		results = append(results, "event: message_delta\ndata: "+string(messageDeltaJSON)+"\n\n")
+		messageDeltaJSON := `{"type":"message_delta","delta":{"stop_reason":"","stop_sequence":null},"usage":{"input_tokens":0,"output_tokens":0}}`
+		messageDeltaJSON, _ = sjson.Set(messageDeltaJSON, "delta.stop_reason", mapOpenAIFinishReasonToAnthropic(param.FinishReason))
+		messageDeltaJSON, _ = sjson.Set(messageDeltaJSON, "usage.input_tokens", inputTokens)
+		messageDeltaJSON, _ = sjson.Set(messageDeltaJSON, "usage.output_tokens", outputTokens)
+		results = append(results, "event: message_delta\ndata: "+messageDeltaJSON+"\n\n")
 		param.MessageDeltaSent = true
 
 		emitMessageStopIfNeeded(param, &results)
@@ -390,12 +321,9 @@ func convertOpenAIDoneToAnthropic(param *ConvertOpenAIResponseToAnthropicParams)
 
 	// Ensure all content blocks are stopped before final events
 	if param.ThinkingContentBlockStarted {
-		contentBlockStop := map[string]interface{}{
-			"type":  "content_block_stop",
-			"index": param.ThinkingContentBlockIndex,
-		}
-		contentBlockStopJSON, _ := json.Marshal(contentBlockStop)
-		results = append(results, "event: content_block_stop\ndata: "+string(contentBlockStopJSON)+"\n\n")
+		contentBlockStopJSON := `{"type":"content_block_stop","index":0}`
+		contentBlockStopJSON, _ = sjson.Set(contentBlockStopJSON, "index", param.ThinkingContentBlockIndex)
+		results = append(results, "event: content_block_stop\ndata: "+contentBlockStopJSON+"\n\n")
 		param.ThinkingContentBlockStarted = false
 		param.ThinkingContentBlockIndex = -1
 	}
@@ -408,24 +336,15 @@ func convertOpenAIDoneToAnthropic(param *ConvertOpenAIResponseToAnthropicParams)
 			blockIndex := param.toolContentBlockIndex(index)
 
 			if accumulator.Arguments.Len() > 0 {
-				inputDelta := map[string]interface{}{
-					"type":  "content_block_delta",
-					"index": blockIndex,
-					"delta": map[string]interface{}{
-						"type":         "input_json_delta",
-						"partial_json": util.FixJSON(accumulator.Arguments.String()),
-					},
-				}
-				inputDeltaJSON, _ := json.Marshal(inputDelta)
-				results = append(results, "event: content_block_delta\ndata: "+string(inputDeltaJSON)+"\n\n")
+				inputDeltaJSON := `{"type":"content_block_delta","index":0,"delta":{"type":"input_json_delta","partial_json":""}}`
+				inputDeltaJSON, _ = sjson.Set(inputDeltaJSON, "index", blockIndex)
+				inputDeltaJSON, _ = sjson.Set(inputDeltaJSON, "delta.partial_json", util.FixJSON(accumulator.Arguments.String()))
+				results = append(results, "event: content_block_delta\ndata: "+inputDeltaJSON+"\n\n")
 			}
 
-			contentBlockStop := map[string]interface{}{
-				"type":  "content_block_stop",
-				"index": blockIndex,
-			}
-			contentBlockStopJSON, _ := json.Marshal(contentBlockStop)
-			results = append(results, "event: content_block_stop\ndata: "+string(contentBlockStopJSON)+"\n\n")
+			contentBlockStopJSON := `{"type":"content_block_stop","index":0}`
+			contentBlockStopJSON, _ = sjson.Set(contentBlockStopJSON, "index", blockIndex)
+			results = append(results, "event: content_block_stop\ndata: "+contentBlockStopJSON+"\n\n")
 			delete(param.ToolCallBlockIndexes, index)
 		}
 		param.ContentBlocksStopped = true
@@ -433,16 +352,9 @@ func convertOpenAIDoneToAnthropic(param *ConvertOpenAIResponseToAnthropicParams)
 
 	// If we haven't sent message_delta yet (no usage info was received), send it now
 	if param.FinishReason != "" && !param.MessageDeltaSent {
-		messageDelta := map[string]interface{}{
-			"type": "message_delta",
-			"delta": map[string]interface{}{
-				"stop_reason":   mapOpenAIFinishReasonToAnthropic(param.FinishReason),
-				"stop_sequence": nil,
-			},
-		}
-
-		messageDeltaJSON, _ := json.Marshal(messageDelta)
-		results = append(results, "event: message_delta\ndata: "+string(messageDeltaJSON)+"\n\n")
+		messageDeltaJSON := `{"type":"message_delta","delta":{"stop_reason":"","stop_sequence":null}}`
+		messageDeltaJSON, _ = sjson.Set(messageDeltaJSON, "delta.stop_reason", mapOpenAIFinishReasonToAnthropic(param.FinishReason))
+		results = append(results, "event: message_delta\ndata: "+messageDeltaJSON+"\n\n")
 		param.MessageDeltaSent = true
 	}
 
@@ -455,105 +367,73 @@ func convertOpenAIDoneToAnthropic(param *ConvertOpenAIResponseToAnthropicParams)
 func convertOpenAINonStreamingToAnthropic(rawJSON []byte) []string {
 	root := gjson.ParseBytes(rawJSON)
 
-	// Build Anthropic response
-	response := map[string]interface{}{
-		"id":            root.Get("id").String(),
-		"type":          "message",
-		"role":          "assistant",
-		"model":         root.Get("model").String(),
-		"content":       []interface{}{},
-		"stop_reason":   nil,
-		"stop_sequence": nil,
-		"usage": map[string]interface{}{
-			"input_tokens":  0,
-			"output_tokens": 0,
-		},
-	}
+	out := `{"id":"","type":"message","role":"assistant","model":"","content":[],"stop_reason":null,"stop_sequence":null,"usage":{"input_tokens":0,"output_tokens":0}}`
+	out, _ = sjson.Set(out, "id", root.Get("id").String())
+	out, _ = sjson.Set(out, "model", root.Get("model").String())
 
 	// Process message content and tool calls
-	var contentBlocks []interface{}
-
-	if choices := root.Get("choices"); choices.Exists() && choices.IsArray() {
+	if choices := root.Get("choices"); choices.Exists() && choices.IsArray() && len(choices.Array()) > 0 {
 		choice := choices.Array()[0] // Take first choice
-		reasoningNode := choice.Get("message.reasoning_content")
-		allReasoning := collectOpenAIReasoningTexts(reasoningNode)
 
-		for _, reasoningText := range allReasoning {
+		reasoningNode := choice.Get("message.reasoning_content")
+		for _, reasoningText := range collectOpenAIReasoningTexts(reasoningNode) {
 			if reasoningText == "" {
 				continue
 			}
-			contentBlocks = append(contentBlocks, map[string]interface{}{
-				"type":     "thinking",
-				"thinking": reasoningText,
-			})
+			block := `{"type":"thinking","thinking":""}`
+			block, _ = sjson.Set(block, "thinking", reasoningText)
+			out, _ = sjson.SetRaw(out, "content.-1", block)
 		}
 
 		// Handle text content
 		if content := choice.Get("message.content"); content.Exists() && content.String() != "" {
-			textBlock := map[string]interface{}{
-				"type": "text",
-				"text": content.String(),
-			}
-			contentBlocks = append(contentBlocks, textBlock)
+			block := `{"type":"text","text":""}`
+			block, _ = sjson.Set(block, "text", content.String())
+			out, _ = sjson.SetRaw(out, "content.-1", block)
 		}
 
 		// Handle tool calls
 		if toolCalls := choice.Get("message.tool_calls"); toolCalls.Exists() && toolCalls.IsArray() {
 			toolCalls.ForEach(func(_, toolCall gjson.Result) bool {
-				toolUseBlock := map[string]interface{}{
-					"type": "tool_use",
-					"id":   toolCall.Get("id").String(),
-					"name": toolCall.Get("function.name").String(),
-				}
+				toolUseBlock := `{"type":"tool_use","id":"","name":"","input":{}}`
+				toolUseBlock, _ = sjson.Set(toolUseBlock, "id", toolCall.Get("id").String())
+				toolUseBlock, _ = sjson.Set(toolUseBlock, "name", toolCall.Get("function.name").String())
 
-				// Parse arguments
-				argsStr := toolCall.Get("function.arguments").String()
-				argsStr = util.FixJSON(argsStr)
-				if argsStr != "" {
-					var args interface{}
-					if err := json.Unmarshal([]byte(argsStr), &args); err == nil {
-						toolUseBlock["input"] = args
+				argsStr := util.FixJSON(toolCall.Get("function.arguments").String())
+				if argsStr != "" && gjson.Valid(argsStr) {
+					argsJSON := gjson.Parse(argsStr)
+					if argsJSON.IsObject() {
+						toolUseBlock, _ = sjson.SetRaw(toolUseBlock, "input", argsJSON.Raw)
 					} else {
-						toolUseBlock["input"] = map[string]interface{}{}
+						toolUseBlock, _ = sjson.SetRaw(toolUseBlock, "input", "{}")
 					}
 				} else {
-					toolUseBlock["input"] = map[string]interface{}{}
+					toolUseBlock, _ = sjson.SetRaw(toolUseBlock, "input", "{}")
 				}
 
-				contentBlocks = append(contentBlocks, toolUseBlock)
+				out, _ = sjson.SetRaw(out, "content.-1", toolUseBlock)
 				return true
 			})
 		}
 
 		// Set stop reason
 		if finishReason := choice.Get("finish_reason"); finishReason.Exists() {
-			response["stop_reason"] = mapOpenAIFinishReasonToAnthropic(finishReason.String())
+			out, _ = sjson.Set(out, "stop_reason", mapOpenAIFinishReasonToAnthropic(finishReason.String()))
 		}
 	}
 
-	response["content"] = contentBlocks
-
 	// Set usage information
 	if usage := root.Get("usage"); usage.Exists() {
-		response["usage"] = map[string]interface{}{
-			"input_tokens":  usage.Get("prompt_tokens").Int(),
-			"output_tokens": usage.Get("completion_tokens").Int(),
-			"reasoning_tokens": func() int64 {
-				if v := usage.Get("completion_tokens_details.reasoning_tokens"); v.Exists() {
-					return v.Int()
-				}
-				return 0
-			}(),
-		}
-	} else {
-		response["usage"] = map[string]interface{}{
-			"input_tokens":  0,
-			"output_tokens": 0,
+		out, _ = sjson.Set(out, "usage.input_tokens", usage.Get("prompt_tokens").Int())
+		out, _ = sjson.Set(out, "usage.output_tokens", usage.Get("completion_tokens").Int())
+		reasoningTokens := int64(0)
+		if v := usage.Get("completion_tokens_details.reasoning_tokens"); v.Exists() {
+			reasoningTokens = v.Int()
 		}
+		out, _ = sjson.Set(out, "usage.reasoning_tokens", reasoningTokens)
 	}
 
-	responseJSON, _ := json.Marshal(response)
-	return []string{string(responseJSON)}
+	return []string{out}
 }
 
 // mapOpenAIFinishReasonToAnthropic maps OpenAI finish reasons to Anthropic equivalents
@@ -600,15 +480,15 @@ func collectOpenAIReasoningTexts(node gjson.Result) []string {
 
 	switch node.Type {
 	case gjson.String:
-		if text := strings.TrimSpace(node.String()); text != "" {
+		if text := node.String(); text != "" {
 			texts = append(texts, text)
 		}
 	case gjson.JSON:
 		if text := node.Get("text"); text.Exists() {
-			if trimmed := strings.TrimSpace(text.String()); trimmed != "" {
-				texts = append(texts, trimmed)
+			if textStr := text.String(); textStr != "" {
+				texts = append(texts, textStr)
 			}
-		} else if raw := strings.TrimSpace(node.Raw); raw != "" && !strings.HasPrefix(raw, "{") && !strings.HasPrefix(raw, "[") {
+		} else if raw := node.Raw; raw != "" && !strings.HasPrefix(raw, "{") && !strings.HasPrefix(raw, "[") {
 			texts = append(texts, raw)
 		}
 	}
@@ -620,12 +500,9 @@ func stopThinkingContentBlock(param *ConvertOpenAIResponseToAnthropicParams, res
 	if !param.ThinkingContentBlockStarted {
 		return
 	}
-	contentBlockStop := map[string]interface{}{
-		"type":  "content_block_stop",
-		"index": param.ThinkingContentBlockIndex,
-	}
-	contentBlockStopJSON, _ := json.Marshal(contentBlockStop)
-	*results = append(*results, "event: content_block_stop\ndata: "+string(contentBlockStopJSON)+"\n\n")
+	contentBlockStopJSON := `{"type":"content_block_stop","index":0}`
+	contentBlockStopJSON, _ = sjson.Set(contentBlockStopJSON, "index", param.ThinkingContentBlockIndex)
+	*results = append(*results, "event: content_block_stop\ndata: "+contentBlockStopJSON+"\n\n")
 	param.ThinkingContentBlockStarted = false
 	param.ThinkingContentBlockIndex = -1
 }
@@ -642,12 +519,9 @@ func stopTextContentBlock(param *ConvertOpenAIResponseToAnthropicParams, results
 	if !param.TextContentBlockStarted {
 		return
 	}
-	contentBlockStop := map[string]interface{}{
-		"type":  "content_block_stop",
-		"index": param.TextContentBlockIndex,
-	}
-	contentBlockStopJSON, _ := json.Marshal(contentBlockStop)
-	*results = append(*results, "event: content_block_stop\ndata: "+string(contentBlockStopJSON)+"\n\n")
+	contentBlockStopJSON := `{"type":"content_block_stop","index":0}`
+	contentBlockStopJSON, _ = sjson.Set(contentBlockStopJSON, "index", param.TextContentBlockIndex)
+	*results = append(*results, "event: content_block_stop\ndata: "+contentBlockStopJSON+"\n\n")
 	param.TextContentBlockStarted = false
 	param.TextContentBlockIndex = -1
 }
@@ -667,29 +541,19 @@ func ConvertOpenAIResponseToClaudeNonStream(_ context.Context, _ string, origina
 	_ = requestRawJSON
 
 	root := gjson.ParseBytes(rawJSON)
+	out := `{"id":"","type":"message","role":"assistant","model":"","content":[],"stop_reason":null,"stop_sequence":null,"usage":{"input_tokens":0,"output_tokens":0}}`
+	out, _ = sjson.Set(out, "id", root.Get("id").String())
+	out, _ = sjson.Set(out, "model", root.Get("model").String())
 
-	response := map[string]interface{}{
-		"id":            root.Get("id").String(),
-		"type":          "message",
-		"role":          "assistant",
-		"model":         root.Get("model").String(),
-		"content":       []interface{}{},
-		"stop_reason":   nil,
-		"stop_sequence": nil,
-		"usage": map[string]interface{}{
-			"input_tokens":  0,
-			"output_tokens": 0,
-		},
-	}
-
-	contentBlocks := make([]interface{}, 0)
 	hasToolCall := false
+	stopReasonSet := false
 
 	if choices := root.Get("choices"); choices.Exists() && choices.IsArray() && len(choices.Array()) > 0 {
 		choice := choices.Array()[0]
 
 		if finishReason := choice.Get("finish_reason"); finishReason.Exists() {
-			response["stop_reason"] = mapOpenAIFinishReasonToAnthropic(finishReason.String())
+			out, _ = sjson.Set(out, "stop_reason", mapOpenAIFinishReasonToAnthropic(finishReason.String()))
+			stopReasonSet = true
 		}
 
 		if message := choice.Get("message"); message.Exists() {
@@ -702,10 +566,9 @@ func ConvertOpenAIResponseToClaudeNonStream(_ context.Context, _ string, origina
 						if textBuilder.Len() == 0 {
 							return
 						}
-						contentBlocks = append(contentBlocks, map[string]interface{}{
-							"type": "text",
-							"text": textBuilder.String(),
-						})
+						block := `{"type":"text","text":""}`
+						block, _ = sjson.Set(block, "text", textBuilder.String())
+						out, _ = sjson.SetRaw(out, "content.-1", block)
 						textBuilder.Reset()
 					}
 
@@ -713,16 +576,14 @@ func ConvertOpenAIResponseToClaudeNonStream(_ context.Context, _ string, origina
 						if thinkingBuilder.Len() == 0 {
 							return
 						}
-						contentBlocks = append(contentBlocks, map[string]interface{}{
-							"type":     "thinking",
-							"thinking": thinkingBuilder.String(),
-						})
+						block := `{"type":"thinking","thinking":""}`
+						block, _ = sjson.Set(block, "thinking", thinkingBuilder.String())
+						out, _ = sjson.SetRaw(out, "content.-1", block)
 						thinkingBuilder.Reset()
 					}
 
 					for _, item := range contentResult.Array() {
-						typeStr := item.Get("type").String()
-						switch typeStr {
+						switch item.Get("type").String() {
 						case "text":
 							flushThinking()
 							textBuilder.WriteString(item.Get("text").String())
@@ -733,25 +594,23 @@ func ConvertOpenAIResponseToClaudeNonStream(_ context.Context, _ string, origina
 							if toolCalls.IsArray() {
 								toolCalls.ForEach(func(_, tc gjson.Result) bool {
 									hasToolCall = true
-									toolUse := map[string]interface{}{
-										"type": "tool_use",
-										"id":   tc.Get("id").String(),
-										"name": tc.Get("function.name").String(),
-									}
+									toolUse := `{"type":"tool_use","id":"","name":"","input":{}}`
+									toolUse, _ = sjson.Set(toolUse, "id", tc.Get("id").String())
+									toolUse, _ = sjson.Set(toolUse, "name", tc.Get("function.name").String())
 
 									argsStr := util.FixJSON(tc.Get("function.arguments").String())
-									if argsStr != "" {
-										var parsed interface{}
-										if err := json.Unmarshal([]byte(argsStr), &parsed); err == nil {
-											toolUse["input"] = parsed
+									if argsStr != "" && gjson.Valid(argsStr) {
+										argsJSON := gjson.Parse(argsStr)
+										if argsJSON.IsObject() {
+											toolUse, _ = sjson.SetRaw(toolUse, "input", argsJSON.Raw)
 										} else {
-											toolUse["input"] = map[string]interface{}{}
+											toolUse, _ = sjson.SetRaw(toolUse, "input", "{}")
 										}
 									} else {
-										toolUse["input"] = map[string]interface{}{}
+										toolUse, _ = sjson.SetRaw(toolUse, "input", "{}")
 									}
 
-									contentBlocks = append(contentBlocks, toolUse)
+									out, _ = sjson.SetRaw(out, "content.-1", toolUse)
 									return true
 								})
 							}
@@ -771,10 +630,9 @@ func ConvertOpenAIResponseToClaudeNonStream(_ context.Context, _ string, origina
 				} else if contentResult.Type == gjson.String {
 					textContent := contentResult.String()
 					if textContent != "" {
-						contentBlocks = append(contentBlocks, map[string]interface{}{
-							"type": "text",
-							"text": textContent,
-						})
+						block := `{"type":"text","text":""}`
+						block, _ = sjson.Set(block, "text", textContent)
+						out, _ = sjson.SetRaw(out, "content.-1", block)
 					}
 				}
 			}
@@ -784,81 +642,52 @@ func ConvertOpenAIResponseToClaudeNonStream(_ context.Context, _ string, origina
 					if reasoningText == "" {
 						continue
 					}
-					contentBlocks = append(contentBlocks, map[string]interface{}{
-						"type":     "thinking",
-						"thinking": reasoningText,
-					})
+					block := `{"type":"thinking","thinking":""}`
+					block, _ = sjson.Set(block, "thinking", reasoningText)
+					out, _ = sjson.SetRaw(out, "content.-1", block)
 				}
 			}
 
 			if toolCalls := message.Get("tool_calls"); toolCalls.Exists() && toolCalls.IsArray() {
 				toolCalls.ForEach(func(_, toolCall gjson.Result) bool {
 					hasToolCall = true
-					toolUseBlock := map[string]interface{}{
-						"type": "tool_use",
-						"id":   toolCall.Get("id").String(),
-						"name": toolCall.Get("function.name").String(),
-					}
+					toolUseBlock := `{"type":"tool_use","id":"","name":"","input":{}}`
+					toolUseBlock, _ = sjson.Set(toolUseBlock, "id", toolCall.Get("id").String())
+					toolUseBlock, _ = sjson.Set(toolUseBlock, "name", toolCall.Get("function.name").String())
 
-					argsStr := toolCall.Get("function.arguments").String()
-					argsStr = util.FixJSON(argsStr)
-					if argsStr != "" {
-						var args interface{}
-						if err := json.Unmarshal([]byte(argsStr), &args); err == nil {
-							toolUseBlock["input"] = args
+					argsStr := util.FixJSON(toolCall.Get("function.arguments").String())
+					if argsStr != "" && gjson.Valid(argsStr) {
+						argsJSON := gjson.Parse(argsStr)
+						if argsJSON.IsObject() {
+							toolUseBlock, _ = sjson.SetRaw(toolUseBlock, "input", argsJSON.Raw)
 						} else {
-							toolUseBlock["input"] = map[string]interface{}{}
+							toolUseBlock, _ = sjson.SetRaw(toolUseBlock, "input", "{}")
 						}
 					} else {
-						toolUseBlock["input"] = map[string]interface{}{}
+						toolUseBlock, _ = sjson.SetRaw(toolUseBlock, "input", "{}")
 					}
 
-					contentBlocks = append(contentBlocks, toolUseBlock)
+					out, _ = sjson.SetRaw(out, "content.-1", toolUseBlock)
 					return true
 				})
 			}
 		}
 	}
 
-	response["content"] = contentBlocks
-
 	if respUsage := root.Get("usage"); respUsage.Exists() {
-		usageJSON := `{}`
-		usageJSON, _ = sjson.Set(usageJSON, "input_tokens", respUsage.Get("prompt_tokens").Int())
-		usageJSON, _ = sjson.Set(usageJSON, "output_tokens", respUsage.Get("completion_tokens").Int())
-		parsedUsage := gjson.Parse(usageJSON).Value().(map[string]interface{})
-		response["usage"] = parsedUsage
-	} else {
-		response["usage"] = `{"input_tokens":0,"output_tokens":0}`
+		out, _ = sjson.Set(out, "usage.input_tokens", respUsage.Get("prompt_tokens").Int())
+		out, _ = sjson.Set(out, "usage.output_tokens", respUsage.Get("completion_tokens").Int())
 	}
 
-	if response["stop_reason"] == nil {
+	if !stopReasonSet {
 		if hasToolCall {
-			response["stop_reason"] = "tool_use"
+			out, _ = sjson.Set(out, "stop_reason", "tool_use")
 		} else {
-			response["stop_reason"] = "end_turn"
+			out, _ = sjson.Set(out, "stop_reason", "end_turn")
 		}
 	}
 
-	if !hasToolCall {
-		if toolBlocks := response["content"].([]interface{}); len(toolBlocks) > 0 {
-			for _, block := range toolBlocks {
-				if m, ok := block.(map[string]interface{}); ok && m["type"] == "tool_use" {
-					hasToolCall = true
-					break
-				}
-			}
-		}
-		if hasToolCall {
-			response["stop_reason"] = "tool_use"
-		}
-	}
-
-	responseJSON, err := json.Marshal(response)
-	if err != nil {
-		return ""
-	}
-	return string(responseJSON)
+	return out
 }
 
 func ClaudeTokenCount(ctx context.Context, count int64) string {

internal/translator/openai/gemini/openai_gemini_request.go

@@ -8,7 +8,6 @@ package gemini
 import (
 	"bytes"
 	"crypto/rand"
-	"encoding/json"
 	"fmt"
 	"math/big"
 	"strings"
@@ -94,7 +93,6 @@ func ConvertGeminiRequestToOpenAI(modelName string, inputRawJSON []byte, stream
 	out, _ = sjson.Set(out, "stream", stream)
 
 	// Process contents (Gemini messages) -> OpenAI messages
-	var openAIMessages []interface{}
 	var toolCallIDs []string // Track tool call IDs for matching with tool results
 
 	// System instruction -> OpenAI system message
@@ -105,22 +103,17 @@ func ConvertGeminiRequestToOpenAI(modelName string, inputRawJSON []byte, stream
 	}
 	if systemInstruction.Exists() {
 		parts := systemInstruction.Get("parts")
-		msg := map[string]interface{}{
-			"role":    "system",
-			"content": []interface{}{},
-		}
-
-		var aggregatedParts []interface{}
+		msg := `{"role":"system","content":[]}`
+		hasContent := false
 
 		if parts.Exists() && parts.IsArray() {
 			parts.ForEach(func(_, part gjson.Result) bool {
 				// Handle text parts
 				if text := part.Get("text"); text.Exists() {
-					formattedText := text.String()
-					aggregatedParts = append(aggregatedParts, map[string]interface{}{
-						"type": "text",
-						"text": formattedText,
-					})
+					contentPart := `{"type":"text","text":""}`
+					contentPart, _ = sjson.Set(contentPart, "text", text.String())
+					msg, _ = sjson.SetRaw(msg, "content.-1", contentPart)
+					hasContent = true
 				}
 
 				// Handle inline data (e.g., images)
@@ -132,20 +125,17 @@ func ConvertGeminiRequestToOpenAI(modelName string, inputRawJSON []byte, stream
 					data := inlineData.Get("data").String()
 					imageURL := fmt.Sprintf("data:%s;base64,%s", mimeType, data)
 
-					aggregatedParts = append(aggregatedParts, map[string]interface{}{
-						"type": "image_url",
-						"image_url": map[string]interface{}{
-							"url": imageURL,
-						},
-					})
+					contentPart := `{"type":"image_url","image_url":{"url":""}}`
+					contentPart, _ = sjson.Set(contentPart, "image_url.url", imageURL)
+					msg, _ = sjson.SetRaw(msg, "content.-1", contentPart)
+					hasContent = true
 				}
 				return true
 			})
 		}
 
-		if len(aggregatedParts) > 0 {
-			msg["content"] = aggregatedParts
-			openAIMessages = append(openAIMessages, msg)
+		if hasContent {
+			out, _ = sjson.SetRaw(out, "messages.-1", msg)
 		}
 	}
 
@@ -159,16 +149,15 @@ func ConvertGeminiRequestToOpenAI(modelName string, inputRawJSON []byte, stream
 				role = "assistant"
 			}
 
-			// Create OpenAI message
-			msg := map[string]interface{}{
-				"role":    role,
-				"content": "",
-			}
+			msg := `{"role":"","content":""}`
+			msg, _ = sjson.Set(msg, "role", role)
 
 			var textBuilder strings.Builder
-			var aggregatedParts []interface{}
+			contentWrapper := `{"arr":[]}`
+			contentPartsCount := 0
 			onlyTextContent := true
-			var toolCalls []interface{}
+			toolCallsWrapper := `{"arr":[]}`
+			toolCallsCount := 0
 
 			if parts.Exists() && parts.IsArray() {
 				parts.ForEach(func(_, part gjson.Result) bool {
@@ -176,10 +165,10 @@ func ConvertGeminiRequestToOpenAI(modelName string, inputRawJSON []byte, stream
 					if text := part.Get("text"); text.Exists() {
 						formattedText := text.String()
 						textBuilder.WriteString(formattedText)
-						aggregatedParts = append(aggregatedParts, map[string]interface{}{
-							"type": "text",
-							"text": formattedText,
-						})
+						contentPart := `{"type":"text","text":""}`
+						contentPart, _ = sjson.Set(contentPart, "text", formattedText)
+						contentWrapper, _ = sjson.SetRaw(contentWrapper, "arr.-1", contentPart)
+						contentPartsCount++
 					}
 
 					// Handle inline data (e.g., images)
@@ -193,12 +182,10 @@ func ConvertGeminiRequestToOpenAI(modelName string, inputRawJSON []byte, stream
 						data := inlineData.Get("data").String()
 						imageURL := fmt.Sprintf("data:%s;base64,%s", mimeType, data)
 
-						aggregatedParts = append(aggregatedParts, map[string]interface{}{
-							"type": "image_url",
-							"image_url": map[string]interface{}{
-								"url": imageURL,
-							},
-						})
+						contentPart := `{"type":"image_url","image_url":{"url":""}}`
+						contentPart, _ = sjson.Set(contentPart, "image_url.url", imageURL)
+						contentWrapper, _ = sjson.SetRaw(contentWrapper, "arr.-1", contentPart)
+						contentPartsCount++
 					}
 
 					// Handle function calls (Gemini) -> tool calls (OpenAI)
@@ -206,44 +193,32 @@ func ConvertGeminiRequestToOpenAI(modelName string, inputRawJSON []byte, stream
 						toolCallID := genToolCallID()
 						toolCallIDs = append(toolCallIDs, toolCallID)
 
-						toolCall := map[string]interface{}{
-							"id":   toolCallID,
-							"type": "function",
-							"function": map[string]interface{}{
-								"name": functionCall.Get("name").String(),
-							},
-						}
+						toolCall := `{"id":"","type":"function","function":{"name":"","arguments":""}}`
+						toolCall, _ = sjson.Set(toolCall, "id", toolCallID)
+						toolCall, _ = sjson.Set(toolCall, "function.name", functionCall.Get("name").String())
 
 						// Convert args to arguments JSON string
 						if args := functionCall.Get("args"); args.Exists() {
-							argsJSON, _ := json.Marshal(args.Value())
-							toolCall["function"].(map[string]interface{})["arguments"] = string(argsJSON)
+							toolCall, _ = sjson.Set(toolCall, "function.arguments", args.Raw)
 						} else {
-							toolCall["function"].(map[string]interface{})["arguments"] = "{}"
+							toolCall, _ = sjson.Set(toolCall, "function.arguments", "{}")
 						}
 
-						toolCalls = append(toolCalls, toolCall)
+						toolCallsWrapper, _ = sjson.SetRaw(toolCallsWrapper, "arr.-1", toolCall)
+						toolCallsCount++
 					}
 
 					// Handle function responses (Gemini) -> tool role messages (OpenAI)
 					if functionResponse := part.Get("functionResponse"); functionResponse.Exists() {
 						// Create tool message for function response
-						toolMsg := map[string]interface{}{
-							"role":         "tool",
-							"tool_call_id": "", // Will be set based on context
-							"content":      "",
-						}
+						toolMsg := `{"role":"tool","tool_call_id":"","content":""}`
 
 						// Convert response.content to JSON string
 						if response := functionResponse.Get("response"); response.Exists() {
-							if content = response.Get("content"); content.Exists() {
-								// Use the content field from the response
-								contentJSON, _ := json.Marshal(content.Value())
-								toolMsg["content"] = string(contentJSON)
+							if contentField := response.Get("content"); contentField.Exists() {
+								toolMsg, _ = sjson.Set(toolMsg, "content", contentField.Raw)
 							} else {
-								// Fallback to entire response
-								responseJSON, _ := json.Marshal(response.Value())
-								toolMsg["content"] = string(responseJSON)
+								toolMsg, _ = sjson.Set(toolMsg, "content", response.Raw)
 							}
 						}
 
@@ -252,13 +227,13 @@ func ConvertGeminiRequestToOpenAI(modelName string, inputRawJSON []byte, stream
 						if len(toolCallIDs) > 0 {
 							// Use the last tool call ID (simple matching by function name)
 							// In a real implementation, you might want more sophisticated matching
-							toolMsg["tool_call_id"] = toolCallIDs[len(toolCallIDs)-1]
+							toolMsg, _ = sjson.Set(toolMsg, "tool_call_id", toolCallIDs[len(toolCallIDs)-1])
 						} else {
 							// Generate a tool call ID if none available
-							toolMsg["tool_call_id"] = genToolCallID()
+							toolMsg, _ = sjson.Set(toolMsg, "tool_call_id", genToolCallID())
 						}
 
-						openAIMessages = append(openAIMessages, toolMsg)
+						out, _ = sjson.SetRaw(out, "messages.-1", toolMsg)
 					}
 
 					return true
@@ -266,170 +241,46 @@ func ConvertGeminiRequestToOpenAI(modelName string, inputRawJSON []byte, stream
 			}
 
 			// Set content
-			if len(aggregatedParts) > 0 {
+			if contentPartsCount > 0 {
 				if onlyTextContent {
-					msg["content"] = textBuilder.String()
+					msg, _ = sjson.Set(msg, "content", textBuilder.String())
 				} else {
-					msg["content"] = aggregatedParts
+					msg, _ = sjson.SetRaw(msg, "content", gjson.Get(contentWrapper, "arr").Raw)
 				}
 			}
 
 			// Set tool calls if any
-			if len(toolCalls) > 0 {
-				msg["tool_calls"] = toolCalls
+			if toolCallsCount > 0 {
+				msg, _ = sjson.SetRaw(msg, "tool_calls", gjson.Get(toolCallsWrapper, "arr").Raw)
 			}
 
-			openAIMessages = append(openAIMessages, msg)
-
-			// switch role {
-			// case "user", "model":
-			// 	// Convert role: model -> assistant
-			// 	if role == "model" {
-			// 		role = "assistant"
-			// 	}
-			//
-			// 	// Create OpenAI message
-			// 	msg := map[string]interface{}{
-			// 		"role":    role,
-			// 		"content": "",
-			// 	}
-			//
-			// 	var contentParts []string
-			// 	var toolCalls []interface{}
-			//
-			// 	if parts.Exists() && parts.IsArray() {
-			// 		parts.ForEach(func(_, part gjson.Result) bool {
-			// 			// Handle text parts
-			// 			if text := part.Get("text"); text.Exists() {
-			// 				contentParts = append(contentParts, text.String())
-			// 			}
-			//
-			// 			// Handle function calls (Gemini) -> tool calls (OpenAI)
-			// 			if functionCall := part.Get("functionCall"); functionCall.Exists() {
-			// 				toolCallID := genToolCallID()
-			// 				toolCallIDs = append(toolCallIDs, toolCallID)
-			//
-			// 				toolCall := map[string]interface{}{
-			// 					"id":   toolCallID,
-			// 					"type": "function",
-			// 					"function": map[string]interface{}{
-			// 						"name": functionCall.Get("name").String(),
-			// 					},
-			// 				}
-			//
-			// 				// Convert args to arguments JSON string
-			// 				if args := functionCall.Get("args"); args.Exists() {
-			// 					argsJSON, _ := json.Marshal(args.Value())
-			// 					toolCall["function"].(map[string]interface{})["arguments"] = string(argsJSON)
-			// 				} else {
-			// 					toolCall["function"].(map[string]interface{})["arguments"] = "{}"
-			// 				}
-			//
-			// 				toolCalls = append(toolCalls, toolCall)
-			// 			}
-			//
-			// 			return true
-			// 		})
-			// 	}
-			//
-			// 	// Set content
-			// 	if len(contentParts) > 0 {
-			// 		msg["content"] = strings.Join(contentParts, "")
-			// 	}
-			//
-			// 	// Set tool calls if any
-			// 	if len(toolCalls) > 0 {
-			// 		msg["tool_calls"] = toolCalls
-			// 	}
-			//
-			// 	openAIMessages = append(openAIMessages, msg)
-			//
-			// case "function":
-			// 	// Handle Gemini function role -> OpenAI tool role
-			// 	if parts.Exists() && parts.IsArray() {
-			// 		parts.ForEach(func(_, part gjson.Result) bool {
-			// 			// Handle function responses (Gemini) -> tool role messages (OpenAI)
-			// 			if functionResponse := part.Get("functionResponse"); functionResponse.Exists() {
-			// 				// Create tool message for function response
-			// 				toolMsg := map[string]interface{}{
-			// 					"role":         "tool",
-			// 					"tool_call_id": "", // Will be set based on context
-			// 					"content":      "",
-			// 				}
-			//
-			// 				// Convert response.content to JSON string
-			// 				if response := functionResponse.Get("response"); response.Exists() {
-			// 					if content = response.Get("content"); content.Exists() {
-			// 						// Use the content field from the response
-			// 						contentJSON, _ := json.Marshal(content.Value())
-			// 						toolMsg["content"] = string(contentJSON)
-			// 					} else {
-			// 						// Fallback to entire response
-			// 						responseJSON, _ := json.Marshal(response.Value())
-			// 						toolMsg["content"] = string(responseJSON)
-			// 					}
-			// 				}
-			//
-			// 				// Try to match with previous tool call ID
-			// 				_ = functionResponse.Get("name").String() // functionName not used for now
-			// 				if len(toolCallIDs) > 0 {
-			// 					// Use the last tool call ID (simple matching by function name)
-			// 					// In a real implementation, you might want more sophisticated matching
-			// 					toolMsg["tool_call_id"] = toolCallIDs[len(toolCallIDs)-1]
-			// 				} else {
-			// 					// Generate a tool call ID if none available
-			// 					toolMsg["tool_call_id"] = genToolCallID()
-			// 				}
-			//
-			// 				openAIMessages = append(openAIMessages, toolMsg)
-			// 			}
-			//
-			// 			return true
-			// 		})
-			// 	}
-			// }
+			out, _ = sjson.SetRaw(out, "messages.-1", msg)
 			return true
 		})
 	}
 
-	// Set messages
-	if len(openAIMessages) > 0 {
-		messagesJSON, _ := json.Marshal(openAIMessages)
-		out, _ = sjson.SetRaw(out, "messages", string(messagesJSON))
-	}
-
 	// Tools mapping: Gemini tools -> OpenAI tools
 	if tools := root.Get("tools"); tools.Exists() && tools.IsArray() {
-		var openAITools []interface{}
 		tools.ForEach(func(_, tool gjson.Result) bool {
 			if functionDeclarations := tool.Get("functionDeclarations"); functionDeclarations.Exists() && functionDeclarations.IsArray() {
 				functionDeclarations.ForEach(func(_, funcDecl gjson.Result) bool {
-					openAITool := map[string]interface{}{
-						"type": "function",
-						"function": map[string]interface{}{
-							"name":        funcDecl.Get("name").String(),
-							"description": funcDecl.Get("description").String(),
-						},
-					}
+					openAITool := `{"type":"function","function":{"name":"","description":""}}`
+					openAITool, _ = sjson.Set(openAITool, "function.name", funcDecl.Get("name").String())
+					openAITool, _ = sjson.Set(openAITool, "function.description", funcDecl.Get("description").String())
 
 					// Convert parameters schema
 					if parameters := funcDecl.Get("parameters"); parameters.Exists() {
-						openAITool["function"].(map[string]interface{})["parameters"] = parameters.Value()
-					} else if parameters = funcDecl.Get("parametersJsonSchema"); parameters.Exists() {
-						openAITool["function"].(map[string]interface{})["parameters"] = parameters.Value()
+						openAITool, _ = sjson.SetRaw(openAITool, "function.parameters", parameters.Raw)
+					} else if parameters := funcDecl.Get("parametersJsonSchema"); parameters.Exists() {
+						openAITool, _ = sjson.SetRaw(openAITool, "function.parameters", parameters.Raw)
 					}
 
-					openAITools = append(openAITools, openAITool)
+					out, _ = sjson.SetRaw(out, "tools.-1", openAITool)
 					return true
 				})
 			}
 			return true
 		})
-
-		if len(openAITools) > 0 {
-			toolsJSON, _ := json.Marshal(openAITools)
-			out, _ = sjson.SetRaw(out, "tools", string(toolsJSON))
-		}
 	}
 
 	// Tool choice mapping (Gemini doesn't have direct equivalent, but we can handle it)

internal/translator/openai/gemini/openai_gemini_response.go

@@ -8,7 +8,6 @@ package gemini
 import (
 	"bytes"
 	"context"
-	"encoding/json"
 	"fmt"
 	"strconv"
 	"strings"
@@ -84,15 +83,12 @@ func ConvertOpenAIResponseToGemini(_ context.Context, _ string, originalRequestR
 					template, _ = sjson.Set(template, "model", model.String())
 				}
 
-				usageObj := map[string]interface{}{
-					"promptTokenCount":     usage.Get("prompt_tokens").Int(),
-					"candidatesTokenCount": usage.Get("completion_tokens").Int(),
-					"totalTokenCount":      usage.Get("total_tokens").Int(),
-				}
+				template, _ = sjson.Set(template, "usageMetadata.promptTokenCount", usage.Get("prompt_tokens").Int())
+				template, _ = sjson.Set(template, "usageMetadata.candidatesTokenCount", usage.Get("completion_tokens").Int())
+				template, _ = sjson.Set(template, "usageMetadata.totalTokenCount", usage.Get("total_tokens").Int())
 				if reasoningTokens := reasoningTokensFromUsage(usage); reasoningTokens > 0 {
-					usageObj["thoughtsTokenCount"] = reasoningTokens
+					template, _ = sjson.Set(template, "usageMetadata.thoughtsTokenCount", reasoningTokens)
 				}
-				template, _ = sjson.Set(template, "usageMetadata", usageObj)
 				return []string{template}
 			}
 			return []string{}
@@ -133,13 +129,8 @@ func ConvertOpenAIResponseToGemini(_ context.Context, _ string, originalRequestR
 						continue
 					}
 					reasoningTemplate := baseTemplate
-					parts := []interface{}{
-						map[string]interface{}{
-							"thought": true,
-							"text":    reasoningText,
-						},
-					}
-					reasoningTemplate, _ = sjson.Set(reasoningTemplate, "candidates.0.content.parts", parts)
+					reasoningTemplate, _ = sjson.Set(reasoningTemplate, "candidates.0.content.parts.0.thought", true)
+					reasoningTemplate, _ = sjson.Set(reasoningTemplate, "candidates.0.content.parts.0.text", reasoningText)
 					chunkOutputs = append(chunkOutputs, reasoningTemplate)
 				}
 			}
@@ -150,13 +141,8 @@ func ConvertOpenAIResponseToGemini(_ context.Context, _ string, originalRequestR
 				(*param).(*ConvertOpenAIResponseToGeminiParams).ContentAccumulator.WriteString(contentText)
 
 				// Create text part for this delta
-				parts := []interface{}{
-					map[string]interface{}{
-						"text": contentText,
-					},
-				}
 				contentTemplate := baseTemplate
-				contentTemplate, _ = sjson.Set(contentTemplate, "candidates.0.content.parts", parts)
+				contentTemplate, _ = sjson.Set(contentTemplate, "candidates.0.content.parts.0.text", contentText)
 				chunkOutputs = append(chunkOutputs, contentTemplate)
 			}
 
@@ -225,24 +211,13 @@ func ConvertOpenAIResponseToGemini(_ context.Context, _ string, originalRequestR
 
 				// If we have accumulated tool calls, output them now
 				if len((*param).(*ConvertOpenAIResponseToGeminiParams).ToolCallsAccumulator) > 0 {
-					var parts []interface{}
+					partIndex := 0
 					for _, accumulator := range (*param).(*ConvertOpenAIResponseToGeminiParams).ToolCallsAccumulator {
-						argsStr := accumulator.Arguments.String()
-						var argsMap map[string]interface{}
-
-						argsMap = parseArgsToMap(argsStr)
-
-						functionCallPart := map[string]interface{}{
-							"functionCall": map[string]interface{}{
-								"name": accumulator.Name,
-								"args": argsMap,
-							},
-						}
-						parts = append(parts, functionCallPart)
-					}
-
-					if len(parts) > 0 {
-						template, _ = sjson.Set(template, "candidates.0.content.parts", parts)
+						namePath := fmt.Sprintf("candidates.0.content.parts.%d.functionCall.name", partIndex)
+						argsPath := fmt.Sprintf("candidates.0.content.parts.%d.functionCall.args", partIndex)
+						template, _ = sjson.Set(template, namePath, accumulator.Name)
+						template, _ = sjson.SetRaw(template, argsPath, parseArgsToObjectRaw(accumulator.Arguments.String()))
+						partIndex++
 					}
 
 					// Clear accumulators
@@ -255,15 +230,12 @@ func ConvertOpenAIResponseToGemini(_ context.Context, _ string, originalRequestR
 
 			// Handle usage information
 			if usage := root.Get("usage"); usage.Exists() {
-				usageObj := map[string]interface{}{
-					"promptTokenCount":     usage.Get("prompt_tokens").Int(),
-					"candidatesTokenCount": usage.Get("completion_tokens").Int(),
-					"totalTokenCount":      usage.Get("total_tokens").Int(),
-				}
+				template, _ = sjson.Set(template, "usageMetadata.promptTokenCount", usage.Get("prompt_tokens").Int())
+				template, _ = sjson.Set(template, "usageMetadata.candidatesTokenCount", usage.Get("completion_tokens").Int())
+				template, _ = sjson.Set(template, "usageMetadata.totalTokenCount", usage.Get("total_tokens").Int())
 				if reasoningTokens := reasoningTokensFromUsage(usage); reasoningTokens > 0 {
-					usageObj["thoughtsTokenCount"] = reasoningTokens
+					template, _ = sjson.Set(template, "usageMetadata.thoughtsTokenCount", reasoningTokens)
 				}
-				template, _ = sjson.Set(template, "usageMetadata", usageObj)
 				results = append(results, template)
 				return true
 			}
@@ -291,46 +263,54 @@ func mapOpenAIFinishReasonToGemini(openAIReason string) string {
 	}
 }
 
-// parseArgsToMap safely parses a JSON string of function arguments into a map.
-// It returns an empty map if the input is empty or cannot be parsed as a JSON object.
-func parseArgsToMap(argsStr string) map[string]interface{} {
+// parseArgsToObjectRaw safely parses a JSON string of function arguments into an object JSON string.
+// It returns "{}" if the input is empty or cannot be parsed as a JSON object.
+func parseArgsToObjectRaw(argsStr string) string {
 	trimmed := strings.TrimSpace(argsStr)
 	if trimmed == "" || trimmed == "{}" {
-		return map[string]interface{}{}
+		return "{}"
 	}
 
 	// First try strict JSON
-	var out map[string]interface{}
-	if errUnmarshal := json.Unmarshal([]byte(trimmed), &out); errUnmarshal == nil {
-		return out
+	if gjson.Valid(trimmed) {
+		strict := gjson.Parse(trimmed)
+		if strict.IsObject() {
+			return strict.Raw
+		}
 	}
 
 	// Tolerant parse: handle streams where values are barewords (e.g., 北京, celsius)
-	tolerant := tolerantParseJSONMap(trimmed)
-	if len(tolerant) > 0 {
+	tolerant := tolerantParseJSONObjectRaw(trimmed)
+	if tolerant != "{}" {
 		return tolerant
 	}
 
 	// Fallback: return empty object when parsing fails
-	return map[string]interface{}{}
+	return "{}"
 }
 
-// tolerantParseJSONMap attempts to parse a JSON-like object string into a map, tolerating
+func escapeSjsonPathKey(key string) string {
+	key = strings.ReplaceAll(key, `\`, `\\`)
+	key = strings.ReplaceAll(key, `.`, `\.`)
+	return key
+}
+
+// tolerantParseJSONObjectRaw attempts to parse a JSON-like object string into a JSON object string, tolerating
 // bareword values (unquoted strings) commonly seen during streamed tool calls.
 // Example input: {"location": 北京, "unit": celsius}
-func tolerantParseJSONMap(s string) map[string]interface{} {
+func tolerantParseJSONObjectRaw(s string) string {
 	// Ensure we operate within the outermost braces if present
 	start := strings.Index(s, "{")
 	end := strings.LastIndex(s, "}")
 	if start == -1 || end == -1 || start >= end {
-		return map[string]interface{}{}
+		return "{}"
 	}
 	content := s[start+1 : end]
 
 	runes := []rune(content)
 	n := len(runes)
 	i := 0
-	result := make(map[string]interface{})
+	result := "{}"
 
 	for i < n {
 		// Skip whitespace and commas
@@ -356,6 +336,7 @@ func tolerantParseJSONMap(s string) map[string]interface{} {
 			break
 		}
 		keyName := jsonStringTokenToRawString(keyToken)
+		sjsonKey := escapeSjsonPathKey(keyName)
 		i = nextIdx
 
 		// Skip whitespace
@@ -375,17 +356,16 @@ func tolerantParseJSONMap(s string) map[string]interface{} {
 		}
 
 		// Parse value (string, number, object/array, bareword)
-		var value interface{}
 		switch runes[i] {
 		case '"':
 			// JSON string
 			valToken, ni := parseJSONStringRunes(runes, i)
 			if ni == -1 {
 				// Malformed; treat as empty string
-				value = ""
+				result, _ = sjson.Set(result, sjsonKey, "")
 				i = n
 			} else {
-				value = jsonStringTokenToRawString(valToken)
+				result, _ = sjson.Set(result, sjsonKey, jsonStringTokenToRawString(valToken))
 				i = ni
 			}
 		case '{', '[':
@@ -394,11 +374,10 @@ func tolerantParseJSONMap(s string) map[string]interface{} {
 			if ni == -1 {
 				i = n
 			} else {
-				var anyVal interface{}
-				if errUnmarshal := json.Unmarshal([]byte(seg), &anyVal); errUnmarshal == nil {
-					value = anyVal
+				if gjson.Valid(seg) {
+					result, _ = sjson.SetRaw(result, sjsonKey, seg)
 				} else {
-					value = seg
+					result, _ = sjson.Set(result, sjsonKey, seg)
 				}
 				i = ni
 			}
@@ -411,21 +390,19 @@ func tolerantParseJSONMap(s string) map[string]interface{} {
 			token := strings.TrimSpace(string(runes[i:j]))
 			// Interpret common JSON atoms and numbers; otherwise treat as string
 			if token == "true" {
-				value = true
+				result, _ = sjson.Set(result, sjsonKey, true)
 			} else if token == "false" {
-				value = false
+				result, _ = sjson.Set(result, sjsonKey, false)
 			} else if token == "null" {
-				value = nil
+				result, _ = sjson.Set(result, sjsonKey, nil)
 			} else if numVal, ok := tryParseNumber(token); ok {
-				value = numVal
+				result, _ = sjson.Set(result, sjsonKey, numVal)
 			} else {
-				value = token
+				result, _ = sjson.Set(result, sjsonKey, token)
 			}
 			i = j
 		}
 
-		result[keyName] = value
-
 		// Skip trailing whitespace and optional comma before next pair
 		for i < n && (runes[i] == ' ' || runes[i] == '\n' || runes[i] == '\r' || runes[i] == '\t') {
 			i++
@@ -463,9 +440,9 @@ func parseJSONStringRunes(runes []rune, start int) (string, int) {
 
 // jsonStringTokenToRawString converts a JSON string token (including quotes) to a raw Go string value.
 func jsonStringTokenToRawString(token string) string {
-	var s string
-	if errUnmarshal := json.Unmarshal([]byte(token), &s); errUnmarshal == nil {
-		return s
+	r := gjson.Parse(token)
+	if r.Type == gjson.String {
+		return r.String()
 	}
 	// Fallback: strip surrounding quotes if present
 	if len(token) >= 2 && token[0] == '"' && token[len(token)-1] == '"' {
@@ -579,7 +556,7 @@ func ConvertOpenAIResponseToGeminiNonStream(_ context.Context, _ string, origina
 				}
 			}
 
-			var parts []interface{}
+			partIndex := 0
 
 			// Handle reasoning content before visible text
 			if reasoning := message.Get("reasoning_content"); reasoning.Exists() {
@@ -587,18 +564,16 @@ func ConvertOpenAIResponseToGeminiNonStream(_ context.Context, _ string, origina
 					if reasoningText == "" {
 						continue
 					}
-					parts = append(parts, map[string]interface{}{
-						"thought": true,
-						"text":    reasoningText,
-					})
+					out, _ = sjson.Set(out, fmt.Sprintf("candidates.0.content.parts.%d.thought", partIndex), true)
+					out, _ = sjson.Set(out, fmt.Sprintf("candidates.0.content.parts.%d.text", partIndex), reasoningText)
+					partIndex++
 				}
 			}
 
 			// Handle content first
 			if content := message.Get("content"); content.Exists() && content.String() != "" {
-				parts = append(parts, map[string]interface{}{
-					"text": content.String(),
-				})
+				out, _ = sjson.Set(out, fmt.Sprintf("candidates.0.content.parts.%d.text", partIndex), content.String())
+				partIndex++
 			}
 
 			// Handle tool calls
@@ -609,27 +584,16 @@ func ConvertOpenAIResponseToGeminiNonStream(_ context.Context, _ string, origina
 						functionName := function.Get("name").String()
 						functionArgs := function.Get("arguments").String()
 
-						// Parse arguments
-						var argsMap map[string]interface{}
-						argsMap = parseArgsToMap(functionArgs)
-
-						functionCallPart := map[string]interface{}{
-							"functionCall": map[string]interface{}{
-								"name": functionName,
-								"args": argsMap,
-							},
-						}
-						parts = append(parts, functionCallPart)
+						namePath := fmt.Sprintf("candidates.0.content.parts.%d.functionCall.name", partIndex)
+						argsPath := fmt.Sprintf("candidates.0.content.parts.%d.functionCall.args", partIndex)
+						out, _ = sjson.Set(out, namePath, functionName)
+						out, _ = sjson.SetRaw(out, argsPath, parseArgsToObjectRaw(functionArgs))
+						partIndex++
 					}
 					return true
 				})
 			}
 
-			// Set parts
-			if len(parts) > 0 {
-				out, _ = sjson.Set(out, "candidates.0.content.parts", parts)
-			}
-
 			// Handle finish reason
 			if finishReason := choice.Get("finish_reason"); finishReason.Exists() {
 				geminiFinishReason := mapOpenAIFinishReasonToGemini(finishReason.String())
@@ -645,15 +609,12 @@ func ConvertOpenAIResponseToGeminiNonStream(_ context.Context, _ string, origina
 
 	// Handle usage information
 	if usage := root.Get("usage"); usage.Exists() {
-		usageObj := map[string]interface{}{
-			"promptTokenCount":     usage.Get("prompt_tokens").Int(),
-			"candidatesTokenCount": usage.Get("completion_tokens").Int(),
-			"totalTokenCount":      usage.Get("total_tokens").Int(),
-		}
+		out, _ = sjson.Set(out, "usageMetadata.promptTokenCount", usage.Get("prompt_tokens").Int())
+		out, _ = sjson.Set(out, "usageMetadata.candidatesTokenCount", usage.Get("completion_tokens").Int())
+		out, _ = sjson.Set(out, "usageMetadata.totalTokenCount", usage.Get("total_tokens").Int())
 		if reasoningTokens := reasoningTokensFromUsage(usage); reasoningTokens > 0 {
-			usageObj["thoughtsTokenCount"] = reasoningTokens
+			out, _ = sjson.Set(out, "usageMetadata.thoughtsTokenCount", reasoningTokens)
 		}
-		out, _ = sjson.Set(out, "usageMetadata", usageObj)
 	}
 
 	return out

internal/translator/openai/openai/responses/openai_openai-responses_response.go

@@ -484,16 +484,12 @@ func ConvertOpenAIChatCompletionsResponseToOpenAIResponses(ctx context.Context,
 					}
 				}
 				// Build response.output using aggregated buffers
-				var outputs []interface{}
+				outputsWrapper := `{"arr":[]}`
 				if st.ReasoningBuf.Len() > 0 {
-					outputs = append(outputs, map[string]interface{}{
-						"id":   st.ReasoningID,
-						"type": "reasoning",
-						"summary": []interface{}{map[string]interface{}{
-							"type": "summary_text",
-							"text": st.ReasoningBuf.String(),
-						}},
-					})
+					item := `{"id":"","type":"reasoning","summary":[{"type":"summary_text","text":""}]}`
+					item, _ = sjson.Set(item, "id", st.ReasoningID)
+					item, _ = sjson.Set(item, "summary.0.text", st.ReasoningBuf.String())
+					outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 				}
 				// Append message items in ascending index order
 				if len(st.MsgItemAdded) > 0 {
@@ -513,18 +509,10 @@ func ConvertOpenAIChatCompletionsResponseToOpenAIResponses(ctx context.Context,
 						if b := st.MsgTextBuf[i]; b != nil {
 							txt = b.String()
 						}
-						outputs = append(outputs, map[string]interface{}{
-							"id":     fmt.Sprintf("msg_%s_%d", st.ResponseID, i),
-							"type":   "message",
-							"status": "completed",
-							"content": []interface{}{map[string]interface{}{
-								"type":        "output_text",
-								"annotations": []interface{}{},
-								"logprobs":    []interface{}{},
-								"text":        txt,
-							}},
-							"role": "assistant",
-						})
+						item := `{"id":"","type":"message","status":"completed","content":[{"type":"output_text","annotations":[],"logprobs":[],"text":""}],"role":"assistant"}`
+						item, _ = sjson.Set(item, "id", fmt.Sprintf("msg_%s_%d", st.ResponseID, i))
+						item, _ = sjson.Set(item, "content.0.text", txt)
+						outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 					}
 				}
 				if len(st.FuncArgsBuf) > 0 {
@@ -547,18 +535,16 @@ func ConvertOpenAIChatCompletionsResponseToOpenAIResponses(ctx context.Context,
 						}
 						callID := st.FuncCallIDs[i]
 						name := st.FuncNames[i]
-						outputs = append(outputs, map[string]interface{}{
-							"id":        fmt.Sprintf("fc_%s", callID),
-							"type":      "function_call",
-							"status":    "completed",
-							"arguments": args,
-							"call_id":   callID,
-							"name":      name,
-						})
+						item := `{"id":"","type":"function_call","status":"completed","arguments":"","call_id":"","name":""}`
+						item, _ = sjson.Set(item, "id", fmt.Sprintf("fc_%s", callID))
+						item, _ = sjson.Set(item, "arguments", args)
+						item, _ = sjson.Set(item, "call_id", callID)
+						item, _ = sjson.Set(item, "name", name)
+						outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 					}
 				}
-				if len(outputs) > 0 {
-					completed, _ = sjson.Set(completed, "response.output", outputs)
+				if gjson.Get(outputsWrapper, "arr.#").Int() > 0 {
+					completed, _ = sjson.SetRaw(completed, "response.output", gjson.Get(outputsWrapper, "arr").Raw)
 				}
 				if st.UsageSeen {
 					completed, _ = sjson.Set(completed, "response.usage.input_tokens", st.PromptTokens)
@@ -681,7 +667,7 @@ func ConvertOpenAIChatCompletionsResponseToOpenAIResponsesNonStream(_ context.Co
 	}
 
 	// Build output list from choices[...]
-	var outputs []interface{}
+	outputsWrapper := `{"arr":[]}`
 	// Detect and capture reasoning content if present
 	rcText := gjson.GetBytes(rawJSON, "choices.0.message.reasoning_content").String()
 	includeReasoning := rcText != ""
@@ -693,21 +679,14 @@ func ConvertOpenAIChatCompletionsResponseToOpenAIResponsesNonStream(_ context.Co
 		if strings.HasPrefix(rid, "resp_") {
 			rid = strings.TrimPrefix(rid, "resp_")
 		}
-		reasoningItem := map[string]interface{}{
-			"id":                fmt.Sprintf("rs_%s", rid),
-			"type":              "reasoning",
-			"encrypted_content": "",
-		}
 		// Prefer summary_text from reasoning_content; encrypted_content is optional
-		var summaries []interface{}
+		reasoningItem := `{"id":"","type":"reasoning","encrypted_content":"","summary":[]}`
+		reasoningItem, _ = sjson.Set(reasoningItem, "id", fmt.Sprintf("rs_%s", rid))
 		if rcText != "" {
-			summaries = append(summaries, map[string]interface{}{
-				"type": "summary_text",
-				"text": rcText,
-			})
+			reasoningItem, _ = sjson.Set(reasoningItem, "summary.0.type", "summary_text")
+			reasoningItem, _ = sjson.Set(reasoningItem, "summary.0.text", rcText)
 		}
-		reasoningItem["summary"] = summaries
-		outputs = append(outputs, reasoningItem)
+		outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", reasoningItem)
 	}
 
 	if choices := root.Get("choices"); choices.Exists() && choices.IsArray() {
@@ -716,18 +695,10 @@ func ConvertOpenAIChatCompletionsResponseToOpenAIResponsesNonStream(_ context.Co
 			if msg.Exists() {
 				// Text message part
 				if c := msg.Get("content"); c.Exists() && c.String() != "" {
-					outputs = append(outputs, map[string]interface{}{
-						"id":     fmt.Sprintf("msg_%s_%d", id, int(choice.Get("index").Int())),
-						"type":   "message",
-						"status": "completed",
-						"content": []interface{}{map[string]interface{}{
-							"type":        "output_text",
-							"annotations": []interface{}{},
-							"logprobs":    []interface{}{},
-							"text":        c.String(),
-						}},
-						"role": "assistant",
-					})
+					item := `{"id":"","type":"message","status":"completed","content":[{"type":"output_text","annotations":[],"logprobs":[],"text":""}],"role":"assistant"}`
+					item, _ = sjson.Set(item, "id", fmt.Sprintf("msg_%s_%d", id, int(choice.Get("index").Int())))
+					item, _ = sjson.Set(item, "content.0.text", c.String())
+					outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 				}
 
 				// Function/tool calls
@@ -736,14 +707,12 @@ func ConvertOpenAIChatCompletionsResponseToOpenAIResponsesNonStream(_ context.Co
 						callID := tc.Get("id").String()
 						name := tc.Get("function.name").String()
 						args := tc.Get("function.arguments").String()
-						outputs = append(outputs, map[string]interface{}{
-							"id":        fmt.Sprintf("fc_%s", callID),
-							"type":      "function_call",
-							"status":    "completed",
-							"arguments": args,
-							"call_id":   callID,
-							"name":      name,
-						})
+						item := `{"id":"","type":"function_call","status":"completed","arguments":"","call_id":"","name":""}`
+						item, _ = sjson.Set(item, "id", fmt.Sprintf("fc_%s", callID))
+						item, _ = sjson.Set(item, "arguments", args)
+						item, _ = sjson.Set(item, "call_id", callID)
+						item, _ = sjson.Set(item, "name", name)
+						outputsWrapper, _ = sjson.SetRaw(outputsWrapper, "arr.-1", item)
 						return true
 					})
 				}
@@ -751,8 +720,8 @@ func ConvertOpenAIChatCompletionsResponseToOpenAIResponsesNonStream(_ context.Co
 			return true
 		})
 	}
-	if len(outputs) > 0 {
-		resp, _ = sjson.Set(resp, "output", outputs)
+	if gjson.Get(outputsWrapper, "arr.#").Int() > 0 {
+		resp, _ = sjson.SetRaw(resp, "output", gjson.Get(outputsWrapper, "arr").Raw)
 	}
 
 	// usage mapping

internal/usage/logger_plugin.go

@@ -6,6 +6,7 @@ package usage
 import (
 	"context"
 	"fmt"
+	"strings"
 	"sync"
 	"sync/atomic"
 	"time"
@@ -90,7 +91,7 @@ type modelStats struct {
 type RequestDetail struct {
 	Timestamp time.Time  `json:"timestamp"`
 	Source    string     `json:"source"`
-	AuthIndex uint64     `json:"auth_index"`
+	AuthIndex string     `json:"auth_index"`
 	Tokens    TokenStats `json:"tokens"`
 	Failed    bool       `json:"failed"`
 }
@@ -281,6 +282,118 @@ func (s *RequestStatistics) Snapshot() StatisticsSnapshot {
 	return result
 }
 
+type MergeResult struct {
+	Added   int64 `json:"added"`
+	Skipped int64 `json:"skipped"`
+}
+
+// MergeSnapshot merges an exported statistics snapshot into the current store.
+// Existing data is preserved and duplicate request details are skipped.
+func (s *RequestStatistics) MergeSnapshot(snapshot StatisticsSnapshot) MergeResult {
+	result := MergeResult{}
+	if s == nil {
+		return result
+	}
+
+	s.mu.Lock()
+	defer s.mu.Unlock()
+
+	seen := make(map[string]struct{})
+	for apiName, stats := range s.apis {
+		if stats == nil {
+			continue
+		}
+		for modelName, modelStatsValue := range stats.Models {
+			if modelStatsValue == nil {
+				continue
+			}
+			for _, detail := range modelStatsValue.Details {
+				seen[dedupKey(apiName, modelName, detail)] = struct{}{}
+			}
+		}
+	}
+
+	for apiName, apiSnapshot := range snapshot.APIs {
+		apiName = strings.TrimSpace(apiName)
+		if apiName == "" {
+			continue
+		}
+		stats, ok := s.apis[apiName]
+		if !ok || stats == nil {
+			stats = &apiStats{Models: make(map[string]*modelStats)}
+			s.apis[apiName] = stats
+		} else if stats.Models == nil {
+			stats.Models = make(map[string]*modelStats)
+		}
+		for modelName, modelSnapshot := range apiSnapshot.Models {
+			modelName = strings.TrimSpace(modelName)
+			if modelName == "" {
+				modelName = "unknown"
+			}
+			for _, detail := range modelSnapshot.Details {
+				detail.Tokens = normaliseTokenStats(detail.Tokens)
+				if detail.Timestamp.IsZero() {
+					detail.Timestamp = time.Now()
+				}
+				key := dedupKey(apiName, modelName, detail)
+				if _, exists := seen[key]; exists {
+					result.Skipped++
+					continue
+				}
+				seen[key] = struct{}{}
+				s.recordImported(apiName, modelName, stats, detail)
+				result.Added++
+			}
+		}
+	}
+
+	return result
+}
+
+func (s *RequestStatistics) recordImported(apiName, modelName string, stats *apiStats, detail RequestDetail) {
+	totalTokens := detail.Tokens.TotalTokens
+	if totalTokens < 0 {
+		totalTokens = 0
+	}
+
+	s.totalRequests++
+	if detail.Failed {
+		s.failureCount++
+	} else {
+		s.successCount++
+	}
+	s.totalTokens += totalTokens
+
+	s.updateAPIStats(stats, modelName, detail)
+
+	dayKey := detail.Timestamp.Format("2006-01-02")
+	hourKey := detail.Timestamp.Hour()
+
+	s.requestsByDay[dayKey]++
+	s.requestsByHour[hourKey]++
+	s.tokensByDay[dayKey] += totalTokens
+	s.tokensByHour[hourKey] += totalTokens
+}
+
+func dedupKey(apiName, modelName string, detail RequestDetail) string {
+	timestamp := detail.Timestamp.UTC().Format(time.RFC3339Nano)
+	tokens := normaliseTokenStats(detail.Tokens)
+	return fmt.Sprintf(
+		"%s|%s|%s|%s|%s|%t|%d|%d|%d|%d|%d",
+		apiName,
+		modelName,
+		timestamp,
+		detail.Source,
+		detail.AuthIndex,
+		detail.Failed,
+		tokens.InputTokens,
+		tokens.OutputTokens,
+		tokens.ReasoningTokens,
+		tokens.CachedTokens,
+		tokens.TotalTokens,
+	)
+}
+
 func resolveAPIIdentifier(ctx context.Context, record coreusage.Record) string {
 	if ctx != nil {
 		if ginCtx, ok := ctx.Value("gin").(*gin.Context); ok && ginCtx != nil {
@@ -340,6 +453,16 @@ func normaliseDetail(detail coreusage.Detail) TokenStats {
 	return tokens
 }
 
+func normaliseTokenStats(tokens TokenStats) TokenStats {
+	if tokens.TotalTokens == 0 {
+		tokens.TotalTokens = tokens.InputTokens + tokens.OutputTokens + tokens.ReasoningTokens
+	}
+	if tokens.TotalTokens == 0 {
+		tokens.TotalTokens = tokens.InputTokens + tokens.OutputTokens + tokens.ReasoningTokens + tokens.CachedTokens
+	}
+	return tokens
+}
+
 func formatHour(hour int) string {
 	if hour < 0 {
 		hour = 0

internal/util/gemini_schema.go

@@ -344,7 +344,7 @@ func cleanupRequiredFields(jsonStr string) string {
 }
 
 // addEmptySchemaPlaceholder adds a placeholder "reason" property to empty object schemas.
-// Claude VALIDATED mode requires at least one property in tool schemas.
+// Claude VALIDATED mode requires at least one required property in tool schemas.
 func addEmptySchemaPlaceholder(jsonStr string) string {
 	// Find all "type" fields
 	paths := findPaths(jsonStr, "type")
@@ -364,6 +364,9 @@ func addEmptySchemaPlaceholder(jsonStr string) string {
 		// Check if properties exists and is empty or missing
 		propsPath := joinPath(parentPath, "properties")
 		propsVal := gjson.Get(jsonStr, propsPath)
+		reqPath := joinPath(parentPath, "required")
+		reqVal := gjson.Get(jsonStr, reqPath)
+		hasRequiredProperties := reqVal.IsArray() && len(reqVal.Array()) > 0
 
 		needsPlaceholder := false
 		if !propsVal.Exists() {
@@ -381,8 +384,22 @@ func addEmptySchemaPlaceholder(jsonStr string) string {
 			jsonStr, _ = sjson.Set(jsonStr, reasonPath+".description", "Brief explanation of why you are calling this tool")
 
 			// Add to required array
-			reqPath := joinPath(parentPath, "required")
 			jsonStr, _ = sjson.Set(jsonStr, reqPath, []string{"reason"})
+			continue
+		}
+
+		// If schema has properties but none are required, add a minimal placeholder.
+		if propsVal.IsObject() && !hasRequiredProperties {
+			// DO NOT add placeholder if it's a top-level schema (parentPath is empty)
+			// or if we've already added a placeholder reason above.
+			if parentPath == "" {
+				continue
+			}
+			placeholderPath := joinPath(propsPath, "_")
+			if !gjson.Get(jsonStr, placeholderPath).Exists() {
+				jsonStr, _ = sjson.Set(jsonStr, placeholderPath+".type", "boolean")
+			}
+			jsonStr, _ = sjson.Set(jsonStr, reqPath, []string{"_"})
 		}
 	}
 

internal/util/gemini_schema_test.go

@@ -127,8 +127,10 @@ func TestCleanJSONSchemaForAntigravity_AnyOfFlattening_SmartSelection(t *testing
 				"type": "object",
 				"description": "Accepts: null | object",
 				"properties": {
+					"_": { "type": "boolean" },
 					"kind": { "type": "string" }
-				}
+				},
+				"required": ["_"]
 			}
 		}
 	}`
@@ -614,71 +616,6 @@ func TestCleanJSONSchemaForAntigravity_MultipleNonNullTypes(t *testing.T) {
 	}
 }
 
-func TestCleanJSONSchemaForGemini_PropertyNamesRemoval(t *testing.T) {
-	// propertyNames is used to validate object property names (e.g., must match a pattern)
-	// Gemini doesn't support this keyword and will reject requests containing it
-	input := `{
-		"type": "object",
-		"properties": {
-			"metadata": {
-				"type": "object",
-				"propertyNames": {
-					"pattern": "^[a-zA-Z_][a-zA-Z0-9_]*$"
-				},
-				"additionalProperties": {
-					"type": "string"
-				}
-			}
-		}
-	}`
-
-	expected := `{
-		"type": "object",
-		"properties": {
-			"metadata": {
-				"type": "object"
-			}
-		}
-	}`
-
-	result := CleanJSONSchemaForGemini(input)
-	compareJSON(t, expected, result)
-
-	// Verify propertyNames is completely removed
-	if strings.Contains(result, "propertyNames") {
-		t.Errorf("propertyNames keyword should be removed, got: %s", result)
-	}
-}
-
-func TestCleanJSONSchemaForGemini_PropertyNamesRemoval_Nested(t *testing.T) {
-	// Test deeply nested propertyNames (as seen in real Claude tool schemas)
-	input := `{
-		"type": "object",
-		"properties": {
-			"items": {
-				"type": "array",
-				"items": {
-					"type": "object",
-					"properties": {
-						"config": {
-							"type": "object",
-							"propertyNames": {
-								"type": "string"
-							}
-						}
-					}
-				}
-			}
-		}
-	}`
-
-	result := CleanJSONSchemaForGemini(input)
-
-	if strings.Contains(result, "propertyNames") {
-		t.Errorf("Nested propertyNames should be removed, got: %s", result)
-	}
-}
-
 func compareJSON(t *testing.T, expectedJSON, actualJSON string) {
 	var expMap, actMap map[string]interface{}
 	errExp := json.Unmarshal([]byte(expectedJSON), &expMap)

internal/util/gemini_thinking.go

@@ -288,37 +288,73 @@ func ApplyDefaultThinkingIfNeeded(model string, body []byte) []byte {
 
 // ApplyGemini3ThinkingLevelFromMetadata applies thinkingLevel from metadata for Gemini 3 models.
 // For standard Gemini API format (generationConfig.thinkingConfig path).
-// This handles the case where reasoning_effort is specified via model name suffix (e.g., model(minimal)).
+// This handles the case where reasoning_effort is specified via model name suffix (e.g., model(minimal))
+// or numeric budget suffix (e.g., model(1000)) which gets converted to a thinkingLevel.
 func ApplyGemini3ThinkingLevelFromMetadata(model string, metadata map[string]any, body []byte) []byte {
-	if !IsGemini3Model(model) {
+	// Use the alias from metadata if available for model type detection
+	lookupModel := ResolveOriginalModel(model, metadata)
+	if !IsGemini3Model(lookupModel) && !IsGemini3Model(model) {
 		return body
 	}
+
+	// Determine which model to use for validation
+	checkModel := model
+	if IsGemini3Model(lookupModel) {
+		checkModel = lookupModel
+	}
+
+	// First try to get effort string from metadata
 	effort, ok := ReasoningEffortFromMetadata(metadata)
-	if !ok || effort == "" {
-		return body
+	if ok && effort != "" {
+		if level, valid := ValidateGemini3ThinkingLevel(checkModel, effort); valid {
+			return ApplyGeminiThinkingLevel(body, level, nil)
+		}
 	}
-	// Validate and apply the thinkingLevel
-	if level, valid := ValidateGemini3ThinkingLevel(model, effort); valid {
-		return ApplyGeminiThinkingLevel(body, level, nil)
+
+	// Fallback: check for numeric budget and convert to thinkingLevel
+	budget, _, _, matched := ThinkingFromMetadata(metadata)
+	if matched && budget != nil {
+		if level, valid := ThinkingBudgetToGemini3Level(checkModel, *budget); valid {
+			return ApplyGeminiThinkingLevel(body, level, nil)
+		}
 	}
+
 	return body
 }
 
 // ApplyGemini3ThinkingLevelFromMetadataCLI applies thinkingLevel from metadata for Gemini 3 models.
 // For Gemini CLI API format (request.generationConfig.thinkingConfig path).
-// This handles the case where reasoning_effort is specified via model name suffix (e.g., model(minimal)).
+// This handles the case where reasoning_effort is specified via model name suffix (e.g., model(minimal))
+// or numeric budget suffix (e.g., model(1000)) which gets converted to a thinkingLevel.
 func ApplyGemini3ThinkingLevelFromMetadataCLI(model string, metadata map[string]any, body []byte) []byte {
-	if !IsGemini3Model(model) {
+	// Use the alias from metadata if available for model type detection
+	lookupModel := ResolveOriginalModel(model, metadata)
+	if !IsGemini3Model(lookupModel) && !IsGemini3Model(model) {
 		return body
 	}
+
+	// Determine which model to use for validation
+	checkModel := model
+	if IsGemini3Model(lookupModel) {
+		checkModel = lookupModel
+	}
+
+	// First try to get effort string from metadata
 	effort, ok := ReasoningEffortFromMetadata(metadata)
-	if !ok || effort == "" {
-		return body
+	if ok && effort != "" {
+		if level, valid := ValidateGemini3ThinkingLevel(checkModel, effort); valid {
+			return ApplyGeminiCLIThinkingLevel(body, level, nil)
+		}
 	}
-	// Validate and apply the thinkingLevel
-	if level, valid := ValidateGemini3ThinkingLevel(model, effort); valid {
-		return ApplyGeminiCLIThinkingLevel(body, level, nil)
+
+	// Fallback: check for numeric budget and convert to thinkingLevel
+	budget, _, _, matched := ThinkingFromMetadata(metadata)
+	if matched && budget != nil {
+		if level, valid := ThinkingBudgetToGemini3Level(checkModel, *budget); valid {
+			return ApplyGeminiCLIThinkingLevel(body, level, nil)
+		}
 	}
+
 	return body
 }
 
@@ -326,15 +362,17 @@ func ApplyGemini3ThinkingLevelFromMetadataCLI(model string, metadata map[string]
 // For Gemini CLI API format (request.generationConfig.thinkingConfig path).
 // Returns the modified body if thinkingConfig was added, otherwise returns the original.
 // For Gemini 3 models, uses thinkingLevel instead of thinkingBudget per Google's documentation.
-func ApplyDefaultThinkingIfNeededCLI(model string, body []byte) []byte {
-	if !ModelHasDefaultThinking(model) {
+func ApplyDefaultThinkingIfNeededCLI(model string, metadata map[string]any, body []byte) []byte {
+	// Use the alias from metadata if available for model property lookup
+	lookupModel := ResolveOriginalModel(model, metadata)
+	if !ModelHasDefaultThinking(lookupModel) && !ModelHasDefaultThinking(model) {
 		return body
 	}
 	if gjson.GetBytes(body, "request.generationConfig.thinkingConfig").Exists() {
 		return body
 	}
 	// Gemini 3 models use thinkingLevel instead of thinkingBudget
-	if IsGemini3Model(model) {
+	if IsGemini3Model(lookupModel) || IsGemini3Model(model) {
 		// Don't set a default - let the API use its dynamic default ("high")
 		// Only set includeThoughts
 		updated, _ := sjson.SetBytes(body, "request.generationConfig.thinkingConfig.includeThoughts", true)

internal/util/sanitize_test.go

@@ -0,0 +1,56 @@
+package util
+
+import (
+	"testing"
+)
+
+func TestSanitizeFunctionName(t *testing.T) {
+	tests := []struct {
+		name     string
+		input    string
+		expected string
+	}{
+		{"Normal", "valid_name", "valid_name"},
+		{"With Dots", "name.with.dots", "name.with.dots"},
+		{"With Colons", "name:with:colons", "name:with:colons"},
+		{"With Dashes", "name-with-dashes", "name-with-dashes"},
+		{"Mixed Allowed", "name.with_dots:colons-dashes", "name.with_dots:colons-dashes"},
+		{"Invalid Characters", "name!with@invalid#chars", "name_with_invalid_chars"},
+		{"Spaces", "name with spaces", "name_with_spaces"},
+		{"Non-ASCII", "name_with_你好_chars", "name_with____chars"},
+		{"Starts with digit", "123name", "_123name"},
+		{"Starts with dot", ".name", "_.name"},
+		{"Starts with colon", ":name", "_:name"},
+		{"Starts with dash", "-name", "_-name"},
+		{"Starts with invalid char", "!name", "_name"},
+		{"Exactly 64 chars", "this_is_a_very_long_name_that_exactly_reaches_sixty_four_charact", "this_is_a_very_long_name_that_exactly_reaches_sixty_four_charact"},
+		{"Too long (65 chars)", "this_is_a_very_long_name_that_exactly_reaches_sixty_four_charactX", "this_is_a_very_long_name_that_exactly_reaches_sixty_four_charact"},
+		{"Very long", "this_is_a_very_long_name_that_exceeds_the_sixty_four_character_limit_for_function_names", "this_is_a_very_long_name_that_exceeds_the_sixty_four_character_l"},
+		{"Starts with digit (64 chars total)", "1234567890123456789012345678901234567890123456789012345678901234", "_123456789012345678901234567890123456789012345678901234567890123"},
+		{"Starts with invalid char (64 chars total)", "!234567890123456789012345678901234567890123456789012345678901234", "_234567890123456789012345678901234567890123456789012345678901234"},
+		{"Empty", "", ""},
+		{"Single character invalid", "@", "_"},
+		{"Single character valid", "a", "a"},
+		{"Single character digit", "1", "_1"},
+		{"Single character underscore", "_", "_"},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got := SanitizeFunctionName(tt.input)
+			if got != tt.expected {
+				t.Errorf("SanitizeFunctionName(%q) = %v, want %v", tt.input, got, tt.expected)
+			}
+			// Verify Gemini compliance
+			if len(got) > 64 {
+				t.Errorf("SanitizeFunctionName(%q) result too long: %d", tt.input, len(got))
+			}
+			if len(got) > 0 {
+				first := got[0]
+				if !((first >= 'a' && first <= 'z') || (first >= 'A' && first <= 'Z') || first == '_') {
+					t.Errorf("SanitizeFunctionName(%q) result starts with invalid char: %c", tt.input, first)
+				}
+			}
+		})
+	}
+}

internal/util/thinking.go

@@ -12,9 +12,18 @@ func ModelSupportsThinking(model string) bool {
 	if model == "" {
 		return false
 	}
+	// First check the global dynamic registry
 	if info := registry.GetGlobalRegistry().GetModelInfo(model); info != nil {
 		return info.Thinking != nil
 	}
+	// Fallback: check static model definitions
+	if info := registry.LookupStaticModelInfo(model); info != nil {
+		return info.Thinking != nil
+	}
+	// Fallback: check Antigravity static config
+	if cfg := registry.GetAntigravityModelConfig()[model]; cfg != nil {
+		return cfg.Thinking != nil
+	}
 	return false
 }
 
@@ -63,11 +72,19 @@ func thinkingRangeFromRegistry(model string) (found bool, min int, max int, zero
 	if model == "" {
 		return false, 0, 0, false, false
 	}
-	info := registry.GetGlobalRegistry().GetModelInfo(model)
-	if info == nil || info.Thinking == nil {
-		return false, 0, 0, false, false
+	// First check global dynamic registry
+	if info := registry.GetGlobalRegistry().GetModelInfo(model); info != nil && info.Thinking != nil {
+		return true, info.Thinking.Min, info.Thinking.Max, info.Thinking.ZeroAllowed, info.Thinking.DynamicAllowed
 	}
-	return true, info.Thinking.Min, info.Thinking.Max, info.Thinking.ZeroAllowed, info.Thinking.DynamicAllowed
+	// Fallback: check static model definitions
+	if info := registry.LookupStaticModelInfo(model); info != nil && info.Thinking != nil {
+		return true, info.Thinking.Min, info.Thinking.Max, info.Thinking.ZeroAllowed, info.Thinking.DynamicAllowed
+	}
+	// Fallback: check Antigravity static config
+	if cfg := registry.GetAntigravityModelConfig()[model]; cfg != nil && cfg.Thinking != nil {
+		return true, cfg.Thinking.Min, cfg.Thinking.Max, cfg.Thinking.ZeroAllowed, cfg.Thinking.DynamicAllowed
+	}
+	return false, 0, 0, false, false
 }
 
 // GetModelThinkingLevels returns the discrete reasoning effort levels for the model.

internal/util/thinking_suffix.go

@@ -7,10 +7,11 @@ import (
 )
 
 const (
-	ThinkingBudgetMetadataKey          = "thinking_budget"
-	ThinkingIncludeThoughtsMetadataKey = "thinking_include_thoughts"
-	ReasoningEffortMetadataKey         = "reasoning_effort"
-	ThinkingOriginalModelMetadataKey   = "thinking_original_model"
+	ThinkingBudgetMetadataKey            = "thinking_budget"
+	ThinkingIncludeThoughtsMetadataKey   = "thinking_include_thoughts"
+	ReasoningEffortMetadataKey           = "reasoning_effort"
+	ThinkingOriginalModelMetadataKey     = "thinking_original_model"
+	ModelMappingOriginalModelMetadataKey = "model_mapping_original_model"
 )
 
 // NormalizeThinkingModel parses dynamic thinking suffixes on model names and returns
@@ -215,6 +216,13 @@ func ResolveOriginalModel(model string, metadata map[string]any) string {
 	}
 
 	if metadata != nil {
+		if v, ok := metadata[ModelMappingOriginalModelMetadataKey]; ok {
+			if s, okStr := v.(string); okStr && strings.TrimSpace(s) != "" {
+				if base := normalize(s); base != "" {
+					return base
+				}
+			}
+		}
 		if v, ok := metadata[ThinkingOriginalModelMetadataKey]; ok {
 			if s, okStr := v.(string); okStr && strings.TrimSpace(s) != "" {
 				if base := normalize(s); base != "" {

internal/util/util.go

@@ -8,12 +8,52 @@ import (
 	"io/fs"
 	"os"
 	"path/filepath"
+	"regexp"
 	"strings"
 
 	"github.com/router-for-me/CLIProxyAPI/v6/internal/config"
 	log "github.com/sirupsen/logrus"
 )
 
+var functionNameSanitizer = regexp.MustCompile(`[^a-zA-Z0-9_.:-]`)
+
+// SanitizeFunctionName ensures a function name matches the requirements for Gemini/Vertex AI.
+// It replaces invalid characters with underscores, ensures it starts with a letter or underscore,
+// and truncates it to 64 characters if necessary.
+// Regex Rule: [^a-zA-Z0-9_.:-] replaced with _.
+func SanitizeFunctionName(name string) string {
+	if name == "" {
+		return ""
+	}
+
+	// Replace invalid characters with underscore
+	sanitized := functionNameSanitizer.ReplaceAllString(name, "_")
+
+	// Ensure it starts with a letter or underscore
+	// Re-reading requirements: Must start with a letter or an underscore.
+	if len(sanitized) > 0 {
+		first := sanitized[0]
+		if !((first >= 'a' && first <= 'z') || (first >= 'A' && first <= 'Z') || first == '_') {
+			// If it starts with an allowed character but not allowed at the beginning (digit, dot, colon, dash),
+			// we must prepend an underscore.
+
+			// To stay within the 64-character limit while prepending, we must truncate first.
+			if len(sanitized) >= 64 {
+				sanitized = sanitized[:63]
+			}
+			sanitized = "_" + sanitized
+		}
+	} else {
+		sanitized = "_"
+	}
+
+	// Truncate to 64 characters
+	if len(sanitized) > 64 {
+		sanitized = sanitized[:64]
+	}
+	return sanitized
+}
+
 // SetLogLevel configures the logrus log level based on the configuration.
 // It sets the log level to DebugLevel if debug mode is enabled, otherwise to InfoLevel.
 func SetLogLevel(cfg *config.Config) {

internal/watcher/config_reload.go

@@ -6,6 +6,7 @@ import (
 	"crypto/sha256"
 	"encoding/hex"
 	"os"
+	"reflect"
 	"time"
 
 	"github.com/router-for-me/CLIProxyAPI/v6/internal/config"
@@ -126,7 +127,7 @@ func (w *Watcher) reloadConfig() bool {
 	}
 
 	authDirChanged := oldConfig == nil || oldConfig.AuthDir != newConfig.AuthDir
-	forceAuthRefresh := oldConfig != nil && oldConfig.ForceModelPrefix != newConfig.ForceModelPrefix
+	forceAuthRefresh := oldConfig != nil && (oldConfig.ForceModelPrefix != newConfig.ForceModelPrefix || !reflect.DeepEqual(oldConfig.OAuthModelMappings, newConfig.OAuthModelMappings))
 
 	log.Infof("config successfully reloaded, triggering client reload")
 	w.reloadClients(authDirChanged, affectedOAuthProviders, forceAuthRefresh)

internal/watcher/diff/config_diff.go

@@ -90,6 +90,11 @@ func BuildConfigChangeDetails(oldCfg, newCfg *config.Config) []string {
 			if !equalStringMap(o.Headers, n.Headers) {
 				changes = append(changes, fmt.Sprintf("gemini[%d].headers: updated", i))
 			}
+			oldModels := SummarizeGeminiModels(o.Models)
+			newModels := SummarizeGeminiModels(n.Models)
+			if oldModels.hash != newModels.hash {
+				changes = append(changes, fmt.Sprintf("gemini[%d].models: updated (%d -> %d entries)", i, oldModels.count, newModels.count))
+			}
 			oldExcluded := SummarizeExcludedModels(o.ExcludedModels)
 			newExcluded := SummarizeExcludedModels(n.ExcludedModels)
 			if oldExcluded.hash != newExcluded.hash {
@@ -120,6 +125,11 @@ func BuildConfigChangeDetails(oldCfg, newCfg *config.Config) []string {
 			if !equalStringMap(o.Headers, n.Headers) {
 				changes = append(changes, fmt.Sprintf("claude[%d].headers: updated", i))
 			}
+			oldModels := SummarizeClaudeModels(o.Models)
+			newModels := SummarizeClaudeModels(n.Models)
+			if oldModels.hash != newModels.hash {
+				changes = append(changes, fmt.Sprintf("claude[%d].models: updated (%d -> %d entries)", i, oldModels.count, newModels.count))
+			}
 			oldExcluded := SummarizeExcludedModels(o.ExcludedModels)
 			newExcluded := SummarizeExcludedModels(n.ExcludedModels)
 			if oldExcluded.hash != newExcluded.hash {
@@ -150,6 +160,11 @@ func BuildConfigChangeDetails(oldCfg, newCfg *config.Config) []string {
 			if !equalStringMap(o.Headers, n.Headers) {
 				changes = append(changes, fmt.Sprintf("codex[%d].headers: updated", i))
 			}
+			oldModels := SummarizeCodexModels(o.Models)
+			newModels := SummarizeCodexModels(n.Models)
+			if oldModels.hash != newModels.hash {
+				changes = append(changes, fmt.Sprintf("codex[%d].models: updated (%d -> %d entries)", i, oldModels.count, newModels.count))
+			}
 			oldExcluded := SummarizeExcludedModels(o.ExcludedModels)
 			newExcluded := SummarizeExcludedModels(n.ExcludedModels)
 			if oldExcluded.hash != newExcluded.hash {
@@ -185,10 +200,18 @@ func BuildConfigChangeDetails(oldCfg, newCfg *config.Config) []string {
 	if oldCfg.AmpCode.ForceModelMappings != newCfg.AmpCode.ForceModelMappings {
 		changes = append(changes, fmt.Sprintf("ampcode.force-model-mappings: %t -> %t", oldCfg.AmpCode.ForceModelMappings, newCfg.AmpCode.ForceModelMappings))
 	}
+	oldUpstreamAPIKeysCount := len(oldCfg.AmpCode.UpstreamAPIKeys)
+	newUpstreamAPIKeysCount := len(newCfg.AmpCode.UpstreamAPIKeys)
+	if !equalUpstreamAPIKeys(oldCfg.AmpCode.UpstreamAPIKeys, newCfg.AmpCode.UpstreamAPIKeys) {
+		changes = append(changes, fmt.Sprintf("ampcode.upstream-api-keys: updated (%d -> %d entries)", oldUpstreamAPIKeysCount, newUpstreamAPIKeysCount))
+	}
 
 	if entries, _ := DiffOAuthExcludedModelChanges(oldCfg.OAuthExcludedModels, newCfg.OAuthExcludedModels); len(entries) > 0 {
 		changes = append(changes, entries...)
 	}
+	if entries, _ := DiffOAuthModelMappingChanges(oldCfg.OAuthModelMappings, newCfg.OAuthModelMappings); len(entries) > 0 {
+		changes = append(changes, entries...)
+	}
 
 	// Remote management (never print the key)
 	if oldCfg.RemoteManagement.AllowRemote != newCfg.RemoteManagement.AllowRemote {
@@ -301,3 +324,43 @@ func formatProxyURL(raw string) string {
 	}
 	return scheme + "://" + host
 }
+
+func equalStringSet(a, b []string) bool {
+	if len(a) == 0 && len(b) == 0 {
+		return true
+	}
+	aSet := make(map[string]struct{}, len(a))
+	for _, k := range a {
+		aSet[strings.TrimSpace(k)] = struct{}{}
+	}
+	bSet := make(map[string]struct{}, len(b))
+	for _, k := range b {
+		bSet[strings.TrimSpace(k)] = struct{}{}
+	}
+	if len(aSet) != len(bSet) {
+		return false
+	}
+	for k := range aSet {
+		if _, ok := bSet[k]; !ok {
+			return false
+		}
+	}
+	return true
+}
+
+// equalUpstreamAPIKeys compares two slices of AmpUpstreamAPIKeyEntry for equality.
+// Comparison is done by count and content (upstream key and client keys).
+func equalUpstreamAPIKeys(a, b []config.AmpUpstreamAPIKeyEntry) bool {
+	if len(a) != len(b) {
+		return false
+	}
+	for i := range a {
+		if strings.TrimSpace(a[i].UpstreamAPIKey) != strings.TrimSpace(b[i].UpstreamAPIKey) {
+			return false
+		}
+		if !equalStringSet(a[i].APIKeys, b[i].APIKeys) {
+			return false
+		}
+	}
+	return true
+}

internal/watcher/diff/model_hash.go

@@ -56,6 +56,36 @@ func ComputeClaudeModelsHash(models []config.ClaudeModel) string {
 	return hashJoined(keys)
 }
 
+// ComputeCodexModelsHash returns a stable hash for Codex model aliases.
+func ComputeCodexModelsHash(models []config.CodexModel) string {
+	keys := normalizeModelPairs(func(out func(key string)) {
+		for _, model := range models {
+			name := strings.TrimSpace(model.Name)
+			alias := strings.TrimSpace(model.Alias)
+			if name == "" && alias == "" {
+				continue
+			}
+			out(strings.ToLower(name) + "|" + strings.ToLower(alias))
+		}
+	})
+	return hashJoined(keys)
+}
+
+// ComputeGeminiModelsHash returns a stable hash for Gemini model aliases.
+func ComputeGeminiModelsHash(models []config.GeminiModel) string {
+	keys := normalizeModelPairs(func(out func(key string)) {
+		for _, model := range models {
+			name := strings.TrimSpace(model.Name)
+			alias := strings.TrimSpace(model.Alias)
+			if name == "" && alias == "" {
+				continue
+			}
+			out(strings.ToLower(name) + "|" + strings.ToLower(alias))
+		}
+	})
+	return hashJoined(keys)
+}
+
 // ComputeExcludedModelsHash returns a normalized hash for excluded model lists.
 func ComputeExcludedModelsHash(excluded []string) string {
 	if len(excluded) == 0 {

internal/watcher/diff/model_hash_test.go

@@ -81,6 +81,15 @@ func TestComputeClaudeModelsHash_Empty(t *testing.T) {
 	}
 }
 
+func TestComputeCodexModelsHash_Empty(t *testing.T) {
+	if got := ComputeCodexModelsHash(nil); got != "" {
+		t.Fatalf("expected empty hash for nil models, got %q", got)
+	}
+	if got := ComputeCodexModelsHash([]config.CodexModel{}); got != "" {
+		t.Fatalf("expected empty hash for empty slice, got %q", got)
+	}
+}
+
 func TestComputeClaudeModelsHash_IgnoresBlankAndDedup(t *testing.T) {
 	a := []config.ClaudeModel{
 		{Name: "m1", Alias: "a1"},
@@ -95,6 +104,20 @@ func TestComputeClaudeModelsHash_IgnoresBlankAndDedup(t *testing.T) {
 	}
 }
 
+func TestComputeCodexModelsHash_IgnoresBlankAndDedup(t *testing.T) {
+	a := []config.CodexModel{
+		{Name: "m1", Alias: "a1"},
+		{Name: " "},
+		{Name: "M1", Alias: "A1"},
+	}
+	b := []config.CodexModel{
+		{Name: "m1", Alias: "a1"},
+	}
+	if h1, h2 := ComputeCodexModelsHash(a), ComputeCodexModelsHash(b); h1 == "" || h1 != h2 {
+		t.Fatalf("expected same hash ignoring blanks/dupes, got %q / %q", h1, h2)
+	}
+}
+
 func TestComputeExcludedModelsHash_Normalizes(t *testing.T) {
 	hash1 := ComputeExcludedModelsHash([]string{" A ", "b", "a"})
 	hash2 := ComputeExcludedModelsHash([]string{"a", " b", "A"})
@@ -157,3 +180,15 @@ func TestComputeClaudeModelsHash_Deterministic(t *testing.T) {
 		t.Fatalf("expected different hash when models change, got %s", h3)
 	}
 }
+
+func TestComputeCodexModelsHash_Deterministic(t *testing.T) {
+	models := []config.CodexModel{{Name: "a", Alias: "A"}, {Name: "b"}}
+	h1 := ComputeCodexModelsHash(models)
+	h2 := ComputeCodexModelsHash(models)
+	if h1 == "" || h1 != h2 {
+		t.Fatalf("expected deterministic hash, got %s / %s", h1, h2)
+	}
+	if h3 := ComputeCodexModelsHash([]config.CodexModel{{Name: "a"}}); h3 == h1 {
+		t.Fatalf("expected different hash when models change, got %s", h3)
+	}
+}

internal/watcher/diff/models_summary.go

@@ -0,0 +1,121 @@
+package diff
+
+import (
+	"crypto/sha256"
+	"encoding/hex"
+	"sort"
+	"strings"
+
+	"github.com/router-for-me/CLIProxyAPI/v6/internal/config"
+)
+
+type GeminiModelsSummary struct {
+	hash  string
+	count int
+}
+
+type ClaudeModelsSummary struct {
+	hash  string
+	count int
+}
+
+type CodexModelsSummary struct {
+	hash  string
+	count int
+}
+
+type VertexModelsSummary struct {
+	hash  string
+	count int
+}
+
+// SummarizeGeminiModels hashes Gemini model aliases for change detection.
+func SummarizeGeminiModels(models []config.GeminiModel) GeminiModelsSummary {
+	if len(models) == 0 {
+		return GeminiModelsSummary{}
+	}
+	keys := normalizeModelPairs(func(out func(key string)) {
+		for _, model := range models {
+			name := strings.TrimSpace(model.Name)
+			alias := strings.TrimSpace(model.Alias)
+			if name == "" && alias == "" {
+				continue
+			}
+			out(strings.ToLower(name) + "|" + strings.ToLower(alias))
+		}
+	})
+	return GeminiModelsSummary{
+		hash:  hashJoined(keys),
+		count: len(keys),
+	}
+}
+
+// SummarizeClaudeModels hashes Claude model aliases for change detection.
+func SummarizeClaudeModels(models []config.ClaudeModel) ClaudeModelsSummary {
+	if len(models) == 0 {
+		return ClaudeModelsSummary{}
+	}
+	keys := normalizeModelPairs(func(out func(key string)) {
+		for _, model := range models {
+			name := strings.TrimSpace(model.Name)
+			alias := strings.TrimSpace(model.Alias)
+			if name == "" && alias == "" {
+				continue
+			}
+			out(strings.ToLower(name) + "|" + strings.ToLower(alias))
+		}
+	})
+	return ClaudeModelsSummary{
+		hash:  hashJoined(keys),
+		count: len(keys),
+	}
+}
+
+// SummarizeCodexModels hashes Codex model aliases for change detection.
+func SummarizeCodexModels(models []config.CodexModel) CodexModelsSummary {
+	if len(models) == 0 {
+		return CodexModelsSummary{}
+	}
+	keys := normalizeModelPairs(func(out func(key string)) {
+		for _, model := range models {
+			name := strings.TrimSpace(model.Name)
+			alias := strings.TrimSpace(model.Alias)
+			if name == "" && alias == "" {
+				continue
+			}
+			out(strings.ToLower(name) + "|" + strings.ToLower(alias))
+		}
+	})
+	return CodexModelsSummary{
+		hash:  hashJoined(keys),
+		count: len(keys),
+	}
+}
+
+// SummarizeVertexModels hashes Vertex-compatible model aliases for change detection.
+func SummarizeVertexModels(models []config.VertexCompatModel) VertexModelsSummary {
+	if len(models) == 0 {
+		return VertexModelsSummary{}
+	}
+	names := make([]string, 0, len(models))
+	for _, model := range models {
+		name := strings.TrimSpace(model.Name)
+		alias := strings.TrimSpace(model.Alias)
+		if name == "" && alias == "" {
+			continue
+		}
+		if alias != "" {
+			name = alias
+		}
+		names = append(names, name)
+	}
+	if len(names) == 0 {
+		return VertexModelsSummary{}
+	}
+	sort.Strings(names)
+	sum := sha256.Sum256([]byte(strings.Join(names, "|")))
+	return VertexModelsSummary{
+		hash:  hex.EncodeToString(sum[:]),
+		count: len(names),
+	}
+}

internal/watcher/diff/oauth_excluded.go

@@ -116,36 +116,3 @@ func SummarizeAmpModelMappings(mappings []config.AmpModelMapping) AmpModelMappin
 		count: len(entries),
 	}
 }
-
-type VertexModelsSummary struct {
-	hash  string
-	count int
-}
-
-// SummarizeVertexModels hashes vertex-compatible models for change detection.
-func SummarizeVertexModels(models []config.VertexCompatModel) VertexModelsSummary {
-	if len(models) == 0 {
-		return VertexModelsSummary{}
-	}
-	names := make([]string, 0, len(models))
-	for _, m := range models {
-		name := strings.TrimSpace(m.Name)
-		alias := strings.TrimSpace(m.Alias)
-		if name == "" && alias == "" {
-			continue
-		}
-		if alias != "" {
-			name = alias
-		}
-		names = append(names, name)
-	}
-	if len(names) == 0 {
-		return VertexModelsSummary{}
-	}
-	sort.Strings(names)
-	sum := sha256.Sum256([]byte(strings.Join(names, "|")))
-	return VertexModelsSummary{
-		hash:  hex.EncodeToString(sum[:]),
-		count: len(names),
-	}
-}

internal/watcher/diff/oauth_model_mappings.go

@@ -0,0 +1,98 @@
+package diff
+
+import (
+	"crypto/sha256"
+	"encoding/hex"
+	"fmt"
+	"sort"
+	"strings"
+
+	"github.com/router-for-me/CLIProxyAPI/v6/internal/config"
+)
+
+type OAuthModelMappingsSummary struct {
+	hash  string
+	count int
+}
+
+// SummarizeOAuthModelMappings summarizes OAuth model mappings per channel.
+func SummarizeOAuthModelMappings(entries map[string][]config.ModelNameMapping) map[string]OAuthModelMappingsSummary {
+	if len(entries) == 0 {
+		return nil
+	}
+	out := make(map[string]OAuthModelMappingsSummary, len(entries))
+	for k, v := range entries {
+		key := strings.ToLower(strings.TrimSpace(k))
+		if key == "" {
+			continue
+		}
+		out[key] = summarizeOAuthModelMappingList(v)
+	}
+	if len(out) == 0 {
+		return nil
+	}
+	return out
+}
+
+// DiffOAuthModelMappingChanges compares OAuth model mappings maps.
+func DiffOAuthModelMappingChanges(oldMap, newMap map[string][]config.ModelNameMapping) ([]string, []string) {
+	oldSummary := SummarizeOAuthModelMappings(oldMap)
+	newSummary := SummarizeOAuthModelMappings(newMap)
+	keys := make(map[string]struct{}, len(oldSummary)+len(newSummary))
+	for k := range oldSummary {
+		keys[k] = struct{}{}
+	}
+	for k := range newSummary {
+		keys[k] = struct{}{}
+	}
+	changes := make([]string, 0, len(keys))
+	affected := make([]string, 0, len(keys))
+	for key := range keys {
+		oldInfo, okOld := oldSummary[key]
+		newInfo, okNew := newSummary[key]
+		switch {
+		case okOld && !okNew:
+			changes = append(changes, fmt.Sprintf("oauth-model-mappings[%s]: removed", key))
+			affected = append(affected, key)
+		case !okOld && okNew:
+			changes = append(changes, fmt.Sprintf("oauth-model-mappings[%s]: added (%d entries)", key, newInfo.count))
+			affected = append(affected, key)
+		case okOld && okNew && oldInfo.hash != newInfo.hash:
+			changes = append(changes, fmt.Sprintf("oauth-model-mappings[%s]: updated (%d -> %d entries)", key, oldInfo.count, newInfo.count))
+			affected = append(affected, key)
+		}
+	}
+	sort.Strings(changes)
+	sort.Strings(affected)
+	return changes, affected
+}
+
+func summarizeOAuthModelMappingList(list []config.ModelNameMapping) OAuthModelMappingsSummary {
+	if len(list) == 0 {
+		return OAuthModelMappingsSummary{}
+	}
+	seen := make(map[string]struct{}, len(list))
+	normalized := make([]string, 0, len(list))
+	for _, mapping := range list {
+		name := strings.ToLower(strings.TrimSpace(mapping.Name))
+		alias := strings.ToLower(strings.TrimSpace(mapping.Alias))
+		if name == "" || alias == "" {
+			continue
+		}
+		key := name + "->" + alias
+		if _, exists := seen[key]; exists {
+			continue
+		}
+		seen[key] = struct{}{}
+		normalized = append(normalized, key)
+	}
+	if len(normalized) == 0 {
+		return OAuthModelMappingsSummary{}
+	}
+	sort.Strings(normalized)
+	sum := sha256.Sum256([]byte(strings.Join(normalized, "|")))
+	return OAuthModelMappingsSummary{
+		hash:  hex.EncodeToString(sum[:]),
+		count: len(normalized),
+	}
+}

internal/watcher/synthesizer/config.go

@@ -66,6 +66,9 @@ func (s *ConfigSynthesizer) synthesizeGeminiKeys(ctx *SynthesisContext) []*corea
 		if base != "" {
 			attrs["base_url"] = base
 		}
+		if hash := diff.ComputeGeminiModelsHash(entry.Models); hash != "" {
+			attrs["models_hash"] = hash
+		}
 		addConfigHeadersToAttrs(entry.Headers, attrs)
 		a := &coreauth.Auth{
 			ID:         id,
@@ -151,6 +154,9 @@ func (s *ConfigSynthesizer) synthesizeCodexKeys(ctx *SynthesisContext) []*coreau
 		if ck.BaseURL != "" {
 			attrs["base_url"] = ck.BaseURL
 		}
+		if hash := diff.ComputeCodexModelsHash(ck.Models); hash != "" {
+			attrs["models_hash"] = hash
+		}
 		addConfigHeadersToAttrs(ck.Headers, attrs)
 		proxyURL := strings.TrimSpace(ck.ProxyURL)
 		a := &coreauth.Auth{

sdk/api/handlers/claude/code_handlers.go

@@ -14,7 +14,6 @@ import (
 	"fmt"
 	"io"
 	"net/http"
-	"time"
 
 	"github.com/gin-gonic/gin"
 	. "github.com/router-for-me/CLIProxyAPI/v6/internal/constant"
@@ -185,14 +184,6 @@ func (h *ClaudeCodeAPIHandler) handleNonStreamingResponse(c *gin.Context, rawJSO
 //   - c: The Gin context for the request.
 //   - rawJSON: The raw JSON request body.
 func (h *ClaudeCodeAPIHandler) handleStreamingResponse(c *gin.Context, rawJSON []byte) {
-	// Set up Server-Sent Events (SSE) headers for streaming response
-	// These headers are essential for maintaining a persistent connection
-	// and enabling real-time streaming of chat completions
-	c.Header("Content-Type", "text/event-stream")
-	c.Header("Cache-Control", "no-cache")
-	c.Header("Connection", "keep-alive")
-	c.Header("Access-Control-Allow-Origin", "*")
-
 	// Get the http.Flusher interface to manually flush the response.
 	// This is crucial for streaming as it allows immediate sending of data chunks
 	flusher, ok := c.Writer.(http.Flusher)
@@ -213,58 +204,82 @@ func (h *ClaudeCodeAPIHandler) handleStreamingResponse(c *gin.Context, rawJSON [
 	cliCtx, cliCancel := h.GetContextWithCancel(h, c, context.Background())
 
 	dataChan, errChan := h.ExecuteStreamWithAuthManager(cliCtx, h.HandlerType(), modelName, rawJSON, "")
-	h.forwardClaudeStream(c, flusher, func(err error) { cliCancel(err) }, dataChan, errChan)
-	return
-}
+	setSSEHeaders := func() {
+		c.Header("Content-Type", "text/event-stream")
+		c.Header("Cache-Control", "no-cache")
+		c.Header("Connection", "keep-alive")
+		c.Header("Access-Control-Allow-Origin", "*")
+	}
 
-func (h *ClaudeCodeAPIHandler) forwardClaudeStream(c *gin.Context, flusher http.Flusher, cancel func(error), data <-chan []byte, errs <-chan *interfaces.ErrorMessage) {
-	// OpenAI-style stream forwarding: write each SSE chunk and flush immediately.
-	// This guarantees clients see incremental output even for small responses.
+	// Peek at the first chunk to determine success or failure before setting headers
 	for {
 		select {
 		case <-c.Request.Context().Done():
-			cancel(c.Request.Context().Err())
+			cliCancel(c.Request.Context().Err())
 			return
-
-		case chunk, ok := <-data:
+		case errMsg, ok := <-errChan:
 			if !ok {
+				// Err channel closed cleanly; wait for data channel.
+				errChan = nil
+				continue
+			}
+			// Upstream failed immediately. Return proper error status and JSON.
+			h.WriteErrorResponse(c, errMsg)
+			if errMsg != nil {
+				cliCancel(errMsg.Error)
+			} else {
+				cliCancel(nil)
+			}
+			return
+		case chunk, ok := <-dataChan:
+			if !ok {
+				// Stream closed without data? Send DONE or just headers.
+				setSSEHeaders()
 				flusher.Flush()
-				cancel(nil)
+				cliCancel(nil)
 				return
 			}
+
+			// Success! Set headers now.
+			setSSEHeaders()
+
+			// Write the first chunk
 			if len(chunk) > 0 {
 				_, _ = c.Writer.Write(chunk)
 				flusher.Flush()
 			}
 
-		case errMsg, ok := <-errs:
-			if !ok {
-				continue
-			}
-			if errMsg != nil {
-				status := http.StatusInternalServerError
-				if errMsg.StatusCode > 0 {
-					status = errMsg.StatusCode
-				}
-				c.Status(status)
-
-				// An error occurred: emit as a proper SSE error event
-				errorBytes, _ := json.Marshal(h.toClaudeError(errMsg))
-				_, _ = fmt.Fprintf(c.Writer, "event: error\ndata: %s\n\n", errorBytes)
-				flusher.Flush()
-			}
-
-			var execErr error
-			if errMsg != nil {
-				execErr = errMsg.Error
-			}
-			cancel(execErr)
+			// Continue streaming the rest
+			h.forwardClaudeStream(c, flusher, func(err error) { cliCancel(err) }, dataChan, errChan)
 			return
-		case <-time.After(500 * time.Millisecond):
 		}
 	}
 }
 
+func (h *ClaudeCodeAPIHandler) forwardClaudeStream(c *gin.Context, flusher http.Flusher, cancel func(error), data <-chan []byte, errs <-chan *interfaces.ErrorMessage) {
+	h.ForwardStream(c, flusher, cancel, data, errs, handlers.StreamForwardOptions{
+		WriteChunk: func(chunk []byte) {
+			if len(chunk) == 0 {
+				return
+			}
+			_, _ = c.Writer.Write(chunk)
+		},
+		WriteTerminalError: func(errMsg *interfaces.ErrorMessage) {
+			if errMsg == nil {
+				return
+			}
+			status := http.StatusInternalServerError
+			if errMsg.StatusCode > 0 {
+				status = errMsg.StatusCode
+			}
+			c.Status(status)
+
+			errorBytes, _ := json.Marshal(h.toClaudeError(errMsg))
+			_, _ = fmt.Fprintf(c.Writer, "event: error\ndata: %s\n\n", errorBytes)
+		},
+	})
+}
+
 type claudeErrorDetail struct {
 	Type    string `json:"type"`
 	Message string `json:"message"`