Files
openhuman/app/src-tauri
91996a1dd0 fix(voice): reduce dictation hallucinations and improve Fn/focus reliability (#385) (#409)
* fix(voice): add per-segment confidence validation in whisper engine (#385)

Reject whisper segments with avg token log-probability below -0.7 or
entropy above 2.4. Return TranscriptionResult with confidence metadata
instead of plain String. Update callers in speech.rs and streaming.rs.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(voice): upgrade default STT model from tiny to base (#385)

Base model produces significantly fewer hallucinations than tiny,
especially in noisy/quiet conditions. User can still override via config.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(voice): add real-time silence gating in audio capture (#385)

Gate sustained silence (>500ms) from being sent to whisper to prevent
hallucinations. Maintain 100ms look-ahead ring buffer so speech onset
after pauses is not clipped. Thresholds adapt to source sample rate.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(voice): fix Fn key timing race condition in hotkey event loop (#385)

start_recording() blocks 1-7s on cpal device init but macOS fires Fn
Release almost immediately, causing skipped cycles. Move recording
start to spawn_blocking so the event loop stays responsive. Buffer
Release events during setup and ensure minimum 1.5s recording duration
when release arrives before recording handle is ready.

Also includes: capture focused app on hotkey press, pass through
pipeline for focus validation before paste.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(voice): validate and restore focus before paste, attempt regardless (#385)

Add expected_app parameter to insert_text(). Before Cmd+V, validate
focus via accessibility API and restore via AppleScript if shifted.
Don't abort paste on focus validation failure — attempt insertion
regardless so text is never silently lost.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

* fix(voice-ui): align voice server RPC response shape in settings panel (#385)

* style(voice): apply rustfmt formatting in text_input

* fix(voice): address CodeRabbit regressions in server, streaming, and settings polling

---------

Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
2026-04-07 09:11:57 -07:00
..
2026-03-29 10:30:18 -07:00
2026-03-29 10:30:18 -07:00
2026-03-29 10:30:18 -07:00
2026-04-07 07:32:01 +00:00
2026-04-07 07:32:01 +00:00