ConversationRelay
A voice agent middleware layer introduced by Twilio. It handles audio transmission and reception, speech recognition, TTS synthesis, and interruption detection, streaming events and text to the backend over WebSocket, and does not maintain conversation history.
Timeline (last 90 days)
Twilio's ConversationRelay is responsible for audio playback and interruption detection, while the model's conversation history is maintained by the application itself.
Twilio recommends that after an interruption occurs, the conversation history must be updated to reflect the speech content that was actually played.
ConversationRelay's interruption message contains two key fields: utteranceUntilInterrupt (the utterance content played up to the interruption) and durationUntilInterruptMs (the duration played).
ConversationRelay itself is not a language model, nor does it maintain conversation history.
ConversationRelay provides backchannel sensitivity control, allowing applications to distinguish backchannel from barge-in via thresholds and durations.
All Facts (5)
Twilio's ConversationRelay is responsible for audio playback and interruption detection, while the model's conversation history is maintained by the application itself.
50%UnverifiedTwilio recommends that after an interruption occurs, the conversation history must be updated to reflect the speech content that was actually played.
50%UnverifiedConversationRelay's interruption message contains two key fields: utteranceUntilInterrupt (the utterance content played up to the interruption) and durationUntilInterruptMs (the duration played).
50%UnverifiedConversationRelay itself is not a language model, nor does it maintain conversation history.
50%UnverifiedConversationRelay provides backchannel sensitivity control, allowing applications to distinguish backchannel from barge-in via thresholds and durations.
50%