diff --git a/src/agents/pi-embedded-subscribe.handlers.tools.media.test.ts b/src/agents/pi-embedded-subscribe.handlers.tools.media.test.ts
index e72e6ed105c..6a607aed011 100644
--- a/src/agents/pi-embedded-subscribe.handlers.tools.media.test.ts
+++ b/src/agents/pi-embedded-subscribe.handlers.tools.media.test.ts
@@ -286,6 +286,35 @@ describe("handleToolExecutionEnd media emission", () => {
     expect(ctx.state.pendingToolAudioAsVoice).toBe(true);
   });
 
+  it("keeps verbose TTS text when structured local media is not trusted", async () => {
+    const ctx = createMockContext({
+      shouldEmitToolOutput: true,
+      onToolResult: vi.fn(),
+      toolResultFormat: "plain",
+      builtinToolNames: new Set(["tts"]),
+    });
+
+    await handleToolExecutionEnd(ctx, {
+      type: "tool_execution_end",
+      toolName: "TTS",
+      toolCallId: "tc-1",
+      isError: false,
+      result: {
+        content: [{ type: "text", text: "(spoken) hello" }],
+        details: {
+          media: {
+            mediaUrl: "/tmp/reply.opus",
+            audioAsVoice: true,
+          },
+        },
+      },
+    });
+
+    expect(ctx.emitToolOutput).toHaveBeenCalled();
+    expect(ctx.state.pendingToolMediaUrls).toEqual([]);
+    expect(ctx.state.pendingToolAudioAsVoice).toBe(false);
+  });
+
   async function handleVerboseGeneratedImage(toolResultFormat: "plain" | "markdown") {
     const ctx = createMockContext({
       shouldEmitToolOutput: true,
diff --git a/src/agents/pi-embedded-subscribe.handlers.tools.ts b/src/agents/pi-embedded-subscribe.handlers.tools.ts
index 0736e3e4de6..2aeb852b59a 100644
--- a/src/agents/pi-embedded-subscribe.handlers.tools.ts
+++ b/src/agents/pi-embedded-subscribe.handlers.tools.ts
@@ -192,9 +192,9 @@ function readApplyPatchSummary(result: unknown): ApplyPatchSummary | null {
 function shouldSuppressStructuredMediaToolOutput(params: {
   toolName: string;
   isToolError: boolean;
-  hasStructuredMedia: boolean;
+  hasDeliverableStructuredMedia: boolean;
 }): boolean {
-  return params.toolName === "tts" && !params.isToolError && params.hasStructuredMedia;
+  return params.toolName === "tts" && !params.isToolError && params.hasDeliverableStructuredMedia;
 }
 
 function buildPatchSummaryText(summary: ApplyPatchSummary): string {
@@ -520,8 +520,16 @@ async function emitToolResultOutput(params: {
   }
 
   const outputText = extractToolResultText(sanitizedResult);
+  const mediaReply = isToolError ? undefined : extractToolResultMediaArtifact(result);
+  const mediaUrls = mediaReply
+    ? filterToolResultMediaUrls(rawToolName, mediaReply.mediaUrls, result, ctx.builtinToolNames)
+    : [];
   const shouldEmitOutput =
-    !shouldSuppressStructuredMediaToolOutput({ toolName, isToolError, hasStructuredMedia }) &&
+    !shouldSuppressStructuredMediaToolOutput({
+      toolName,
+      isToolError,
+      hasDeliverableStructuredMedia: hasStructuredMedia && mediaUrls.length > 0,
+    }) &&
     (ctx.shouldEmitToolOutput() || shouldEmitCompactToolOutput({ toolName, result, outputText }));
   if (shouldEmitOutput) {
     if (outputText) {
@@ -543,16 +551,9 @@ async function emitToolResultOutput(params: {
     return;
   }
 
-  const mediaReply = extractToolResultMediaArtifact(result);
   if (!mediaReply) {
     return;
   }
-  const mediaUrls = filterToolResultMediaUrls(
-    rawToolName,
-    mediaReply.mediaUrls,
-    result,
-    ctx.builtinToolNames,
-  );
   const pendingMediaUrls =
     emittedToolOutputMediaUrls.length === 0
       ? mediaUrls