justLV 44c7be03b8 Only persist assistant history after successful TTS delivery ... Move the conversational backend's _finalize() call out of stream()'s finally block and expose it as a public commit(text) method instead. The previous behavior persisted whatever was yielded even when the stream errored, the user interrupted, or a TTS send failed — so saved history diverged from what the user actually heard, and the next turn replayed phantom context to the LLM. main.py now calls backend.commit(response_text) only after a turn successfully completes and produced content. Agentic backend gets a no-op commit() since history lives on the remote service.		2026-04-12 19:11:09 -07:00
..
conversation	Only persist assistant history after successful TTS delivery	2026-04-12 19:11:09 -07:00
services	Pre-publish fixes: local TTS key, multicast crash guard, doc drift	2026-04-12 19:09:58 -07:00
__init__.py	Add modular async pipeline server and ESP32 mDNS fallback	2026-02-07 15:04:12 -08:00
audio.py	End-of-speech protocol, LED tweaks, call-end guard	2026-04-07 16:41:59 -07:00
config.yaml.example	Refine stall classifier prompt and group benchmark cases by label	2026-04-12 19:08:40 -07:00
device.py	Stream agentic LLM responses, add contextual stall classifier, rename backends	2026-04-12 13:55:59 -07:00
main.py	Only persist assistant history after successful TTS delivery	2026-04-12 19:11:09 -07:00
protocol.py	Simplify device state model, reduce TCP churn, fix multi-device identity	2026-04-09 13:32:23 -07:00
vad.py	Replace webrtcvad with Silero VAD (ONNX, no PyTorch)	2026-02-07 17:00:02 -08:00