api has llm response event

examples/mic_client.py

@@ -4,10 +4,12 @@ Microphone client for testing duplex voice conversation.
 
 This client captures audio from the microphone, sends it to the server,
 and plays back the AI's voice response through the speakers.
+It also displays the LLM's text responses in the console.
 
 Usage:
     python examples/mic_client.py --url ws://localhost:8000/ws
     python examples/mic_client.py --url ws://localhost:8000/ws --chat "Hello!"
+    python examples/mic_client.py --url ws://localhost:8000/ws --verbose
 
 Requirements:
     pip install sounddevice soundfile websockets numpy
@@ -101,6 +103,9 @@ class MicrophoneClient:
         # Interrupt handling - discard audio until next trackStart
         self._discard_audio = False
         self._audio_sequence = 0  # Track audio sequence to detect stale chunks
+        
+        # Verbose mode for streaming LLM responses
+        self.verbose = False
     
     async def connect(self) -> None:
         """Connect to WebSocket server."""
@@ -314,6 +319,17 @@ class MicrophoneClient:
             # Server-side TTFB event
             latency_ms = event.get("latencyMs", 0)
             print(f"← [TTFB] Server reported latency: {latency_ms}ms")
+        elif event_type == "llmResponse":
+            # LLM text response
+            text = event.get("text", "")
+            is_final = event.get("isFinal", False)
+            if is_final:
+                # Print final LLM response
+                print(f"← AI: {text}")
+            elif self.verbose:
+                # Show streaming chunks only in verbose mode
+                display_text = text[:60] + "..." if len(text) > 60 else text
+                print(f"  [streaming] {display_text}")
         elif event_type == "trackStart":
             print("← Bot started speaking")
             # IMPORTANT: Accept audio again after trackStart
@@ -552,6 +568,11 @@ async def main():
         action="store_true",
         help="Disable interactive mode"
     )
+    parser.add_argument(
+        "--verbose", "-v",
+        action="store_true",
+        help="Show streaming LLM response chunks"
+    )
     
     args = parser.parse_args()
     
@@ -565,6 +586,7 @@ async def main():
         input_device=args.input_device,
         output_device=args.output_device
     )
+    client.verbose = args.verbose
     
     await client.run(
         chat_message=args.chat,