diff --git a/CHANGELOG.md b/CHANGELOG.md
index 9d05f853c8f..4c7729f34da 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -81,6 +81,7 @@ Docs: https://docs.openclaw.ai
 - Web UI/Assistant text: strip internal `<relevant-memories>...</relevant-memories>` scaffolding from rendered assistant messages (while preserving code-fence literals), preventing memory-context leakage in chat output for models that echo internal blocks. (#29851) Thanks @Valkster70.
 - Dashboard/Sessions: allow authenticated Control UI clients to delete and patch sessions while still blocking regular webchat clients from session mutation RPCs, fixing Dashboard session delete failures. (#21264) Thanks @jskoiz.
 - TUI/Session model status: clear stale runtime model identity when model overrides change so `/model` updates are reflected immediately in `sessions.patch` responses and `sessions.list` status surfaces. (#28619) Thanks @lejean2000.
+- Memory/Hybrid recall: when strict hybrid scoring yields no hits, preserve keyword-backed matches using a text-weight floor so freshly indexed lexical canaries no longer disappear behind `minScore` filtering. (#29112) Thanks @ceo-nada.
 - Podman/Quadlet setup: fix `sed` escaping and UID mismatch in Podman Quadlet setup. (#26414) Thanks @KnHack and @vincentkoc.
 - Browser/Navigate: resolve the correct `targetId` in navigate responses after renderer swaps. (#25326) Thanks @stone-jin and @vincentkoc.
 - Agents/Ollama discovery: skip Ollama discovery when explicit models are configured. (#28827) Thanks @Kansodata and @vincentkoc.
diff --git a/src/memory/index.test.ts b/src/memory/index.test.ts
index 9174805105d..861862d4f5c 100644
--- a/src/memory/index.test.ts
+++ b/src/memory/index.test.ts
@@ -98,6 +98,7 @@ describe("memory index", () => {
     model?: string;
     vectorEnabled?: boolean;
     cacheEnabled?: boolean;
+    minScore?: number;
     hybrid?: { enabled: boolean; vectorWeight?: number; textWeight?: number };
   }): TestCfg {
     return {
@@ -112,7 +113,7 @@ describe("memory index", () => {
             chunking: { tokens: 4000, overlap: 0 },
             sync: { watch: false, onSessionStart: false, onSearch: true },
             query: {
-              minScore: 0,
+              minScore: params.minScore ?? 0,
               hybrid: params.hybrid ?? { enabled: false },
             },
             cache: params.cacheEnabled ? { enabled: true } : undefined,
@@ -367,6 +368,25 @@ describe("memory index", () => {
     expect(results[0]?.path).toContain("memory/2026-01-12.md");
   });
 
+  it("preserves keyword-only hybrid hits when minScore exceeds text weight", async () => {
+    const cfg = createCfg({
+      storePath: indexMainPath,
+      minScore: 0.35,
+      hybrid: { enabled: true, vectorWeight: 0.7, textWeight: 0.3 },
+    });
+    const manager = await getPersistentManager(cfg);
+
+    const status = manager.status();
+    if (!status.fts?.available) {
+      return;
+    }
+
+    await manager.sync({ reason: "test" });
+    const results = await manager.search("zebra");
+    expect(results.length).toBeGreaterThan(0);
+    expect(results[0]?.path).toContain("memory/2026-01-12.md");
+  });
+
   it("reports vector availability after probe", async () => {
     const cfg = createCfg({ storePath: indexVectorPath, vectorEnabled: true });
     const manager = await getPersistentManager(cfg);
diff --git a/src/memory/manager.ts b/src/memory/manager.ts
index 708e9e7b2c7..36460df87ad 100644
--- a/src/memory/manager.ts
+++ b/src/memory/manager.ts
@@ -311,8 +311,28 @@ export class MemoryIndexManager extends MemoryManagerEmbeddingOps implements Mem
       mmr: hybrid.mmr,
       temporalDecay: hybrid.temporalDecay,
     });
+    const strict = merged.filter((entry) => entry.score >= minScore);
+    if (strict.length > 0 || keywordResults.length === 0) {
+      return strict.slice(0, maxResults);
+    }
 
-    return merged.filter((entry) => entry.score >= minScore).slice(0, maxResults);
+    // Hybrid defaults can produce keyword-only matches with max score equal to
+    // textWeight (for example 0.3). If minScore is higher (for example 0.35),
+    // these exact lexical hits get filtered out even when they are the only
+    // relevant results.
+    const relaxedMinScore = Math.min(minScore, hybrid.textWeight);
+    const keywordKeys = new Set(
+      keywordResults.map(
+        (entry) => `${entry.source}:${entry.path}:${entry.startLine}:${entry.endLine}`,
+      ),
+    );
+    return merged
+      .filter(
+        (entry) =>
+          keywordKeys.has(`${entry.source}:${entry.path}:${entry.startLine}:${entry.endLine}`) &&
+          entry.score >= relaxedMinScore,
+      )
+      .slice(0, maxResults);
   }
 
   private async searchVector(