修复一些已知问题

api/app/features/conversation/ws/pipeline.py

@@ -539,7 +539,7 @@ async def process_user_message(
         is_from_voice = bool(segment.audio_url)
         voice_session_id = _voice_session_id_from_audio_url(segment.audio_url)
         audio_dur = getattr(segment, "audio_duration_seconds", None)
-        responses = await chat_orchestrator.process_user_message(
+        turn = await chat_orchestrator.process_user_message(
             conversation_id=conversation_id,
             user_message=user_message,
             user=user,
@@ -553,6 +553,8 @@ async def process_user_message(
             user_message_timestamp=user_message_timestamp,
             audio_duration_seconds=audio_dur,
         )
+        responses = turn.messages
+        skip_tts = turn.skip_tts
 
         segment.agent_response = "\n\n".join(responses)
         _mark_conversation_active(conversation)
@@ -574,12 +576,14 @@ async def process_user_message(
                     "timestamp": datetime.now(timezone.utc).isoformat(),
                 },
             )
-            url = await _send_tts_audio(
-                conversation_id,
-                response_text,
-                chunk_index=i,
-                chunk_total=n,
-            )
+            url = None
+            if not skip_tts:
+                url = await _send_tts_audio(
+                    conversation_id,
+                    response_text,
+                    chunk_index=i,
+                    chunk_total=n,
+                )
             if url:
                 tts_urls.append(url)
             if i < n - 1: