Sentinel value pushed onto the QueueAudioOutput stream when a segment of agent
speech has finished (i.e. flush() was called). Consumers should treat it as an
end-of-segment marker and emit the corresponding "speech ended" signal on their
downstream protocol (websocket, RPC, etc.).
Sentinel value pushed onto the QueueAudioOutput stream when a segment of agent speech has finished (i.e.
flush()was called). Consumers should treat it as an end-of-segment marker and emit the corresponding "speech ended" signal on their downstream protocol (websocket, RPC, etc.).Ref: python livekit-agents/livekit/agents/voice/avatar/_types.py (AudioSegmentEnd)