pipecat · Issues· 341 open

Feature request: expose provider response/request ids (OpenAI Responses, Deepgram, ElevenLabs) without DEBUG logging

Question: recovering when an LLM stream starts but never reaches a sentence end (slow-streaming window)

TelnyxFrameSerializer silently drops OutputTransportMessageFrame — every other telephony serializer passes it through

Interrupting the bot right as it calls a tool can leave the tool call stuck "in progress"

Twilio/Telnyx/Exotel serializers crash the whole call on a malformed WebSocket message

TTS usage metrics are lost for TTSSpeakFrame text in TOKEN-streaming mode: InterruptionFrame discards _streamed_text before it can flush

feat: add `should_summarize_callback` param to LLMContextSummarizer

Feature request: safe opt-in empty-user-turn recovery via a developer message and one LLM run

FunctionCallResultFrame arriving while the user is still speaking is silently dropped, the LLM is never re-invoked and nothing retries

Is there a documented recovery when an LLM stream starts but stalls before a sentence completes?

Is PipelineWorker.remove_observer() supported for an observer to detach itself from inside its own event handler?

DeepgramFluxSTTService: a blocked websocket send (peer stops reading, socket in CLOSE_WAIT) stalls the whole pipeline with no error, no reconnect and no log line

TranscriptionFrame queued behind the frame that starts a user turn is dropped by that turn's own interruption

No public way to settle outstanding function calls before `on_pipeline_finished` runs

TTSService with push_text_frames=False drops the LLMFullResponseEndFrame of any turn that produces no audio — tool-only turns never end downstream