FE-1570: Optimize spoken Brunch responses for Voice mode - #9496

Closed

kostandinang wants to merge 15 commits into

mainfrom

kostandin/fe-1570-voice-optimized-brunch-responses

Commits on Sep 2, 2026

Separate Voice context from canonical questions

kostandinang
and
ampagent
committed
9ad0fa5
Prepare concise interview speech with Realtime

kostandinang
and
ampagent
committed
1c08e05
Speak prepared context with exact Brunch questions

kostandinang
and
ampagent
committed
7d0c78f
Keep prepared Voice responses lifecycle-safe

kostandinang
and
ampagent
committed
19e5c43
Add canonical Voice replay actions

kostandinang
and
ampagent
committed
73f1aec
Adapt Voice replay fixtures to mute state

kostandinang
and
cursoragent
committed
d864f14
Document Realtime-prepared Brunch responses

kostandinang
and
ampagent
committed
b604287
Clarify Voice composer availability

kostandinang
and
ampagent
committed
26a2526
Mark Voice preparation as experimental

kostandinang
and
ampagent
committed
4827b39
Fix Voice preparation and replay races

kostandinang
and
ampagent
committed
ab6d6fe
Correlate Voice replay with speech completion

kostandinang
and
ampagent
committed
7318293
Complete canceled Realtime function calls

kostandinang
and
cursoragent
committed
5ea7e9c
Bind Voice replay to one response lifecycle

kostandinang
and
ampagent
committed
1fe880e
Prevent stale Voice context and replay

kostandinang
and
cursoragent
committed
943dfc2
Release Voice replay after transcription

kostandinang
and
cursoragent
committed
c7fe8a2