Repository Issue Activity (beta)

#735 Global Shortcut Key not working on macOS 26

#928 Paste targets stale window when dictating twice without leaving Screen Sharing.app

#927 Custom enhancement model on Azure gpt-5.x: test passes, real enhancement fails with HTTP 400 (hard-coded temperature 0.3)

#923 VoiceInk Refine: prompt-steerability + chained Refine→prompted-pass mode?

#885 Transcription model requests

#919 Support / interest in Nix or alternative packaging for local builds

#918 App icon looks out of place

#530 docs: include more troubleshooting guide in the building.md

#916 Recording produces 100% silent audio (0 chars transcribed) unless macOS Dictation is running simultaneously

#915 Audio degraded in-app before transcription: per-file peak normalization + un-filtered downsampling hurt speech intelligibility

#914 Occasional Unexpected Language Switch (French / English)

#911 Copy transcription directly from the History list without expanding the row

#908 Scribe V2 long recordings fail due to 30-second ElevenLabs timeout

#899 Intermittent transcription failure: "Audio file is too short" after recording

#910 Add macOS Accessibility API (AX) context with OCR fallback

#907 Custom enhancement models cannot express required request protocols (S1-mini)

#905 Differentiate between left and right command and alt keys

#771 Enhancement failed: Enhancement request timed out. Check your connection or increase the timeout duration.

#895 [Bug]: Deepgram streaming converts spoken number words into digits due to hardcoded numerals=true

#894 Gemini transcription takes very long — request sends no thinkingConfig, so gemini-3.7-flash thinks at default level

#898 Recorder panel stops appearing after sleep/wake — recording still works, only a relaunch brings the bar back (2.11)

#687 Intermittent empty or truncated transcription after hotkey trigger

#893 Whisper initial prompt doesn't follow the mode's language — non-English modes drift into English/romanized output

#892 Transcription Failed: Invalid audio data provided. Must be at least 300ms of 16kHz audio.

#896 Refine parameters for custom LLMs