wx44wx/pipecat
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update docstring

Browse Source
This commit is contained in:
James Hush
2026-02-12 12:30:10 +08:00
parent eb0ce5aea1
commit 9786c4f8da
1 changed files with 6 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
src/pipecat/turns/user_stop/speech_timeout_user_turn_stop_strategy.py
View File

@@ -34,8 +34,12 @@ class SpeechTimeoutUserTurnStopStrategy(BaseUserTurnStopStrategy):
after the user stops speaking, adjusted by the VAD stop_secs.
For services that support finalization (TranscriptionFrame.finalized=True),
the turn can be triggered immediately once the finalized transcript is
received and the user resume speaking timeout has elapsed.
receiving the finalized transcript allows the strategy to shorten the
timeout by removing the STT wait component, since only the
`user_speech_timeout` portion is still needed. If `user_speech_timeout`
has already elapsed when the transcript arrives, the original timeout
continues running to provide a buffer for VAD to detect any resumed
speech before triggering.
"""
def __init__(self, *, user_speech_timeout: float = 0.6, **kwargs):