Merge pull request #3298 from pipecat-ai/aleix/push-user-started-speaking-first

push UserStartedSpeakingFrame before interruption

examples/foundational/07c-interruptible-deepgram-vad.py

@@ -104,7 +104,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
 
     @stt.event_handler("on_speech_started")
     async def on_speech_started(stt, *args, **kwargs):
-        await task.queue_frames([InterruptionFrame(), UserStartedSpeakingFrame()])
+        await task.queue_frames([UserStartedSpeakingFrame(), InterruptionFrame()])
 
     @stt.event_handler("on_utterance_end")
     async def on_utterance_end(stt, *args, **kwargs):

src/pipecat/services/deepgram/flux/stt.py

@@ -591,8 +591,8 @@ class DeepgramFluxSTTService(WebsocketSTTService):
         """
         logger.debug("User started speaking")
         self._user_is_speaking = True
-        await self.push_interruption_task_frame_and_wait()
         await self.broadcast_frame(UserStartedSpeakingFrame)
+        await self.push_interruption_task_frame_and_wait()
         await self.start_metrics()
         await self._call_event_handler("on_start_of_turn", transcript)
         if transcript:

src/pipecat/services/grok/realtime/llm.py

@@ -652,8 +652,8 @@ class GrokRealtimeLLMService(LLMService):
     async def _handle_evt_speech_started(self, evt):
         """Handle speech started event from VAD."""
         await self._truncate_current_audio_response()
+        await self.broadcast_frame(UserStartedSpeakingFrame)
         await self.push_interruption_task_frame_and_wait()
-        await self.push_frame(UserStartedSpeakingFrame())
 
     async def _handle_evt_speech_stopped(self, evt):
         """Handle speech stopped event from VAD."""

src/pipecat/services/openai/realtime/llm.py

@@ -735,8 +735,8 @@ class OpenAIRealtimeLLMService(LLMService):
 
     async def _handle_evt_speech_started(self, evt):
         await self._truncate_current_audio_response()
+        await self.broadcast_frame(UserStartedSpeakingFrame)
         await self.push_interruption_task_frame_and_wait()
-        await self.push_frame(UserStartedSpeakingFrame())
 
     async def _handle_evt_speech_stopped(self, evt):
         await self.start_ttfb_metrics()

src/pipecat/services/openai_realtime_beta/openai.py

@@ -656,8 +656,8 @@ class OpenAIRealtimeBetaLLMService(LLMService):
 
     async def _handle_evt_speech_started(self, evt):
         await self._truncate_current_audio_response()
+        await self.broadcast_frame(UserStartedSpeakingFrame)
         await self.push_interruption_task_frame_and_wait()
-        await self.push_frame(UserStartedSpeakingFrame())
 
     async def _handle_evt_speech_stopped(self, evt):
         await self.start_ttfb_metrics()

src/pipecat/services/speechmatics/stt.py

@@ -604,8 +604,8 @@ class SpeechmaticsSTTService(STTService):
             message: the message payload.
         """
         logger.debug(f"{self} StartOfTurn received")
-        await self.push_interruption_task_frame_and_wait()
         await self.broadcast_frame(UserStartedSpeakingFrame)
+        await self.push_interruption_task_frame_and_wait()
         # await self.start_processing_metrics()
 
     async def _handle_end_of_turn(self, message: dict[str, Any]) -> None: