Add google stt voice activity timeout

[AhmadIbrahiim]

Add VoiceActivityTimeout support to Google STT V2

Fixes #4804

Adds timeout controls for speech detection in Google STT.

What's new

Two new parameters to control turn detection timing:

• speech_start_timeout - How long to wait for speech to begin
• speech_end_timeout - How much silence before considering the turn complete

Both support fractional seconds and handle the protobuf Duration conversion automatically.

Why?

Without these, you can't tune how patient the STT is with pauses or how long it waits for users to start talking. This helps prevent awkward cutoffs or long dead air.

Real-world impact: Chirp 3 has ~2.4s latency on short utterances like "hi". With speech_end_timeout=0.5, this can be reduced to ~500-800ms.

Usage

stt = google.STT(
    model="chirp_3",
    speech_start_timeout=10.0,
    speech_end_timeout=0.5,  # 500ms silence = end of turn
)

# Update on the fly
stt.update_options(speech_end_timeout=1.0)

Testing

13 tests covering defaults, updates, fractional seconds, and V1/V2 compatibility. All passing.

[CLAassistant]

All committers have signed the CLA.

[AhmadIbrahiim]

Fixes #4804

[AhmadIbrahiim]

Hey @theomonnom @longcw @keepingitneil - would appreciate a review on this when you get a chance!

This adds voice_activity_timeout support to Google STT which is critical for real-time voice agents. Without it, Chirp 3 has ~2.4s latency on short utterances making it impractical for conversational AI.

See issue #4804 for details. Thanks! 🙏

[devin-ai-integration]

Devin Review found 1 potential issue.

View 4 additional findings in Devin Review.