Merge pull request #1128 from golbin/whisper-api

Add Whisper STT service using OpenAI API

examples/foundational/07g-interruptible-openai.py

@@ -18,7 +18,7 @@ from pipecat.pipeline.pipeline import Pipeline
 from pipecat.pipeline.runner import PipelineRunner
 from pipecat.pipeline.task import PipelineParams, PipelineTask
 from pipecat.processors.aggregators.openai_llm_context import OpenAILLMContext
-from pipecat.services.openai import OpenAILLMService, OpenAITTSService
+from pipecat.services.openai import OpenAILLMService, OpenAISTTService, OpenAITTSService
 from pipecat.transports.services.daily import DailyParams, DailyTransport
 
 load_dotenv(override=True)
@@ -37,12 +37,22 @@ async def main():
             "Respond bot",
             DailyParams(
                 audio_out_enabled=True,
-                transcription_enabled=True,
+                audio_out_sample_rate=24000,
+                transcription_enabled=False,
                 vad_enabled=True,
                 vad_analyzer=SileroVADAnalyzer(),
+                vad_audio_passthrough=True,
             ),
         )
 
+        # You can use the OpenAI compatible API like Groq.
+        # stt = OpenAISTTService(
+        #     base_url="https://api.groq.com/openai/v1",
+        #     api_key="gsk_***",
+        #     model="whisper-large-v3",
+        # )
+        stt = OpenAISTTService(api_key=os.getenv("OPENAI_API_KEY"), model="whisper-1")
+
         tts = OpenAITTSService(api_key=os.getenv("OPENAI_API_KEY"), voice="alloy")
 
         llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"), model="gpt-4o")
@@ -60,6 +70,7 @@ async def main():
         pipeline = Pipeline(
             [
                 transport.input(),  # Transport user input
+                stt,  # STT
                 context_aggregator.user(),  # User responses
                 llm,  # LLM
                 tts,  # TTS

src/pipecat/services/openai.py

@@ -30,6 +30,7 @@ from pipecat.frames.frames import (
     OpenAILLMContextAssistantTimestampFrame,
     StartFrame,
     StartInterruptionFrame,
+    TranscriptionFrame,
     TTSAudioRawFrame,
     TTSStartedFrame,
     TTSStoppedFrame,
@@ -48,7 +49,12 @@ from pipecat.processors.aggregators.openai_llm_context import (
     OpenAILLMContextFrame,
 )
 from pipecat.processors.frame_processor import FrameDirection
-from pipecat.services.ai_services import ImageGenService, LLMService, TTSService
+from pipecat.services.ai_services import (
+    ImageGenService,
+    LLMService,
+    SegmentedSTTService,
+    TTSService,
+)
 from pipecat.utils.time import time_now_iso8601
 
 try:
@@ -59,6 +65,7 @@ try:
         BadRequestError,
         DefaultAsyncHttpxClient,
     )
+    from openai.types.audio import Transcription
     from openai.types.chat import ChatCompletionChunk, ChatCompletionMessageParam
 except ModuleNotFoundError as e:
     logger.error(f"Exception: {e}")
@@ -391,6 +398,61 @@ class OpenAIImageGenService(ImageGenService):
             yield frame
 
 
+class OpenAISTTService(SegmentedSTTService):
+    """OpenAI Speech-to-Text (STT) service.
+
+    This service uses OpenAI's Whisper API to convert audio to text.
+
+    Args:
+        model: Whisper model to use. Defaults to "whisper-1".
+        api_key: OpenAI API key. Defaults to None.
+        base_url: API base URL. Defaults to None.
+        **kwargs: Additional arguments passed to SegmentedSTTService.
+    """
+
+    def __init__(
+        self,
+        *,
+        model: str = "whisper-1",
+        api_key: Optional[str] = None,
+        base_url: Optional[str] = None,
+        **kwargs,
+    ):
+        super().__init__(**kwargs)
+        self.set_model_name(model)
+        self._client = AsyncOpenAI(api_key=api_key, base_url=base_url)
+
+    async def set_model(self, model: str):
+        self.set_model_name(model)
+
+    def can_generate_metrics(self) -> bool:
+        return True
+
+    async def run_stt(self, audio: bytes) -> AsyncGenerator[Frame, None]:
+        try:
+            await self.start_processing_metrics()
+            await self.start_ttfb_metrics()
+
+            response: Transcription = await self._client.audio.transcriptions.create(
+                file=("audio.wav", audio, "audio/wav"), model=self.model_name
+            )
+
+            await self.stop_ttfb_metrics()
+            await self.stop_processing_metrics()
+
+            text = response.text.strip()
+
+            if text:
+                logger.debug(f"Transcription: [{text}]")
+                yield TranscriptionFrame(text, "", time_now_iso8601())
+            else:
+                logger.warning("Received empty transcription from API")
+
+        except Exception as e:
+            logger.exception(f"Exception during transcription: {e}")
+            yield ErrorFrame(f"Error during transcription: {str(e)}")
+
+
 class OpenAITTSService(TTSService):
     """OpenAI Text-to-Speech service that generates audio from text.