Merge branch 'router-for-me:main' into main

fix(translators): handle gjson string types in Claude request processing to ensure consistent content parsing

config.example.yaml

@@ -39,6 +39,9 @@ api-keys:
 # 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)

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/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 {
+	if !cfg.CommercialMode {
-		requestLogger = optionState.requestLoggerFactory(cfg, configFilePath)
+		if optionState.requestLoggerFactory != nil {
-	}
+			requestLogger = optionState.requestLoggerFactory(cfg, configFilePath)
-	if requestLogger != nil {
+		}
-		engine.Use(middleware.RequestLoggingMiddleware(requestLogger))
+		if requestLogger != nil {
-		if setter, ok := requestLogger.(interface{ SetEnabled(bool) }); ok {
+			engine.Use(middleware.RequestLoggingMiddleware(requestLogger))
-			toggle = setter.SetEnabled
+			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)
@@ -538,6 +542,7 @@ 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)

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"`
 

internal/logging/gin_logger.go

@@ -73,17 +73,15 @@ func GinLogrusLogger() gin.HandlerFunc {
 		method := c.Request.Method
 		errorMessage := c.Errors.ByType(gin.ErrorTypePrivate).String()
 
+		if requestID == "" {
+			requestID = "--------"
+		}
 		logLine := fmt.Sprintf("%3d | %13v | %15s | %-7s \"%s\"", statusCode, latency, clientIP, method, path)
 		if errorMessage != "" {
 			logLine = logLine + " | " + errorMessage
 		}
 
-		var entry *log.Entry
+		entry := log.WithField("request_id", requestID)
-		if requestID != "" {
-			entry = log.WithField("request_id", requestID)
-		} else {
-			entry = log.WithField("request_id", "--------")
-		}
 
 		switch {
 		case statusCode >= http.StatusInternalServerError:

internal/logging/global_logger.go

@@ -40,25 +40,22 @@ 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")
 
-	reqID := ""
+	reqID := "--------"
 	if id, ok := entry.Data["request_id"].(string); ok && id != "" {
 		reqID = id
 	}
 
-	callerFile := "unknown"
+	level := entry.Level.String()
-	callerLine := 0
+	if level == "warning" {
-	if entry.Caller != nil {
+		level = "warn"
-		callerFile = filepath.Base(entry.Caller.File)
-		callerLine = entry.Caller.Line
 	}
-
+	levelStr := fmt.Sprintf("%-5s", level)
-	levelStr := fmt.Sprintf("%-5s", entry.Level.String())
 
 	var formatted string
-	if reqID != "" {
+	if entry.Caller != nil {
-		formatted = fmt.Sprintf("[%s] [%s] [%s:%d] | %s | %s\n", timestamp, levelStr, callerFile, callerLine, reqID, message)
+		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:%d] %s\n", timestamp, levelStr, callerFile, callerLine, message)
+		formatted = fmt.Sprintf("[%s] [%s] [%s] %s\n", timestamp, reqID, levelStr, message)
 	}
 	buffer.WriteString(formatted)
 

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},
@@ -740,6 +741,7 @@ func GetIFlowModels() []*ModelInfo {
 		{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-m2.1", DisplayName: "MiniMax-M2.1", Description: "MiniMax M2.1", Created: 1766448000},
 	}
 	models := make([]*ModelInfo, 0, len(entries))
 	for _, entry := range entries {

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{
+	return parseGeminiFamilyUsageDetail(node)
-		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
 }
 
 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{
+	return parseGeminiFamilyUsageDetail(node)
-		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
 }
 
 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{
+	return parseGeminiFamilyUsageDetail(node), true
-		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
 }
 
 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{
+	return parseGeminiFamilyUsageDetail(node), true
-		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
 }
 
 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{
+	return parseGeminiFamilyUsageDetail(node)
-		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
 }
 
 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{
+	return parseGeminiFamilyUsageDetail(node), true
-		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
 }
 
 var stopChunkWithoutUsage sync.Map

internal/translator/antigravity/claude/antigravity_claude_response.go

@@ -271,11 +271,11 @@ 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()
-		params.CachedTokenCount = usageResult.Get("cachedContentTokenCount").Int()
 		if params.CandidatesTokenCount == 0 && params.TotalTokenCount > 0 {
 			params.CandidatesTokenCount = params.TotalTokenCount - params.PromptTokenCount - params.ThoughtsTokenCount
 			if params.CandidatesTokenCount < 0 {

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

@@ -249,8 +249,28 @@ func ConvertOpenAIRequestToAntigravity(modelName string, inputRawJSON []byte, _
 				p := 0
 				if content.Type == gjson.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
@@ -305,6 +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)
 				}
 			}
 		}

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

@@ -87,15 +87,15 @@ 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()
-		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)
@@ -181,12 +181,14 @@ func ConvertAntigravityResponseToOpenAI(_ context.Context, _ string, originalReq
 					mimeType = "image/png"
 				}
 				imageURL := fmt.Sprintf("data:%s;base64,%s", mimeType, data)
-				imagePayload := `{"image_url":{"url":""},"type":"image_url"}`
-				imagePayload, _ = sjson.Set(imagePayload, "image_url.url", imageURL)
 				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", imagePayload)
 			}

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

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

@@ -218,8 +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
@@ -260,6 +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)
 				}
 			}
 		}

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

@@ -170,12 +170,14 @@ func ConvertCliResponseToOpenAI(_ context.Context, _ string, originalRequestRawJ
 					mimeType = "image/png"
 				}
 				imageURL := fmt.Sprintf("data:%s;base64,%s", mimeType, data)
-				imagePayload := `{"image_url":{"url":""},"type":"image_url"}`
-				imagePayload, _ = sjson.Set(imagePayload, "image_url.url", imageURL)
 				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", imagePayload)
 			}

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

@@ -233,18 +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.
@@ -261,7 +258,6 @@ func ConvertOpenAIRequestToGemini(modelName string, inputRawJSON []byte, _ bool)
 							}
 						}
 					}
-					out, _ = sjson.SetRawBytes(out, "contents.-1", node)
 				}
 
 				// Tool calls -> single model content with functionCall parts
@@ -302,6 +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)
 				}
 			}
 		}

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

@@ -89,15 +89,15 @@ 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()
-		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)
@@ -182,12 +182,14 @@ func ConvertGeminiResponseToOpenAI(_ context.Context, _ string, originalRequestR
 					mimeType = "image/png"
 				}
 				imageURL := fmt.Sprintf("data:%s;base64,%s", mimeType, data)
-				imagePayload := `{"image_url":{"url":""},"type":"image_url"}`
-				imagePayload, _ = sjson.Set(imagePayload, "image_url.url", imageURL)
 				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", imagePayload)
 			}
@@ -316,12 +318,14 @@ func ConvertGeminiResponseToOpenAINonStream(_ context.Context, _ string, origina
 					mimeType = "image/png"
 				}
 				imageURL := fmt.Sprintf("data:%s;base64,%s", mimeType, data)
-				imagePayload := `{"image_url":{"url":""},"type":"image_url"}`
-				imagePayload, _ = sjson.Set(imagePayload, "image_url.url", imageURL)
 				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", imagePayload)
 			}

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
+	// Track pending tool results that should be attached to the next user message
-	toolCallIDToName := make(map[string]string)
+	// This is critical for LiteLLM-translated requests where tool results appear
-	for _, msg := range messagesArray {
+	// as separate "tool" role messages between assistant and user messages
-		if msg.Get("role").String() == "assistant" {
+	var pendingToolResults []KiroToolResult
-			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
-						}
-					}
-				}
-			}
-		}
-	}
 
 	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
+				// Collect pending tool results to attach to the next user message
-				// For now, collect them and they'll be handled when we build the current message
+				pendingToolResults = append(pendingToolResults, toolResult)
-				currentToolResults = append(currentToolResults, 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/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

sdk/api/handlers/handlers_stream_bootstrap_test.go

@@ -99,7 +99,7 @@ func TestExecuteStreamWithAuthManager_RetriesBeforeFirstByte(t *testing.T) {
 		Streaming: sdkconfig.StreamingConfig{
 			BootstrapRetries: &bootstrapRetries,
 		},
-	}, manager, nil)
+	}, manager)
 	dataChan, errChan := handler.ExecuteStreamWithAuthManager(context.Background(), "openai", "test-model", []byte(`{"model":"test-model"}`), "")
 	if dataChan == nil || errChan == nil {
 		t.Fatalf("expected non-nil channels")

sdk/cliproxy/auth/conductor.go

@@ -203,10 +203,10 @@ func (m *Manager) Register(ctx context.Context, auth *Auth) (*Auth, error) {
 	if auth == nil {
 		return nil, nil
 	}
-	auth.EnsureIndex()
 	if auth.ID == "" {
 		auth.ID = uuid.NewString()
 	}
+	auth.EnsureIndex()
 	m.mu.Lock()
 	m.auths[auth.ID] = auth.Clone()
 	m.mu.Unlock()
@@ -221,7 +221,7 @@ func (m *Manager) Update(ctx context.Context, auth *Auth) (*Auth, error) {
 		return nil, nil
 	}
 	m.mu.Lock()
-	if existing, ok := m.auths[auth.ID]; ok && existing != nil && !auth.indexAssigned && auth.Index == 0 {
+	if existing, ok := m.auths[auth.ID]; ok && existing != nil && !auth.indexAssigned && auth.Index == "" {
 		auth.Index = existing.Index
 		auth.indexAssigned = existing.indexAssigned
 	}

sdk/cliproxy/auth/types.go

@@ -1,11 +1,12 @@
 package auth
 
 import (
+	"crypto/sha256"
+	"encoding/hex"
 	"encoding/json"
 	"strconv"
 	"strings"
 	"sync"
-	"sync/atomic"
 	"time"
 
 	baseauth "github.com/router-for-me/CLIProxyAPI/v6/internal/auth"
@@ -15,8 +16,8 @@ import (
 type Auth struct {
 	// ID uniquely identifies the auth record across restarts.
 	ID string `json:"id"`
-	// Index is a monotonically increasing runtime identifier used for diagnostics.
+	// Index is a stable runtime identifier derived from auth metadata (not persisted).
-	Index uint64 `json:"-"`
+	Index string `json:"-"`
 	// Provider is the upstream provider key (e.g. "gemini", "claude").
 	Provider string `json:"provider"`
 	// Prefix optionally namespaces models for routing (e.g., "teamA/gemini-3-pro-preview").
@@ -94,12 +95,6 @@ type ModelState struct {
 	UpdatedAt time.Time `json:"updated_at"`
 }
 
-var authIndexCounter atomic.Uint64
-
-func nextAuthIndex() uint64 {
-	return authIndexCounter.Add(1) - 1
-}
-
 // Clone shallow copies the Auth structure, duplicating maps to avoid accidental mutation.
 func (a *Auth) Clone() *Auth {
 	if a == nil {
@@ -128,15 +123,41 @@ func (a *Auth) Clone() *Auth {
 	return &copyAuth
 }
 
-// EnsureIndex returns the global index, assigning one if it was not set yet.
+func stableAuthIndex(seed string) string {
-func (a *Auth) EnsureIndex() uint64 {
+	seed = strings.TrimSpace(seed)
-	if a == nil {
+	if seed == "" {
-		return 0
+		return ""
 	}
-	if a.indexAssigned {
+	sum := sha256.Sum256([]byte(seed))
+	return hex.EncodeToString(sum[:8])
+}
+
+// EnsureIndex returns a stable index derived from the auth file name or API key.
+func (a *Auth) EnsureIndex() string {
+	if a == nil {
+		return ""
+	}
+	if a.indexAssigned && a.Index != "" {
 		return a.Index
 	}
-	idx := nextAuthIndex()
+
+	seed := strings.TrimSpace(a.FileName)
+	if seed != "" {
+		seed = "file:" + seed
+	} else if a.Attributes != nil {
+		if apiKey := strings.TrimSpace(a.Attributes["api_key"]); apiKey != "" {
+			seed = "api_key:" + apiKey
+		}
+	}
+	if seed == "" {
+		if id := strings.TrimSpace(a.ID); id != "" {
+			seed = "id:" + id
+		} else {
+			return ""
+		}
+	}
+
+	idx := stableAuthIndex(seed)
 	a.Index = idx
 	a.indexAssigned = true
 	return idx

sdk/cliproxy/usage/manager.go

@@ -14,7 +14,7 @@ type Record struct {
 	Model       string
 	APIKey      string
 	AuthID      string
-	AuthIndex   uint64
+	AuthIndex   string
 	Source      string
 	RequestedAt time.Time
 	Failed      bool