Repository navigation
Troubleshooting
| Problem | Fix |
|---|---|
| macOS: "app can't be opened" | Early builds aren't notarized. Go to System Settings → Privacy & Security → Open Anyway, or right-click → Open, or run xattr -dr com.apple.quarantine "/Applications/Local Transcribe.app". |
| Windows SmartScreen warning | More info → Run anyway. Builds aren't code-signed yet. |
| Linux: app won't start | Install webkit2gtk-4.1 (e.g. sudo apt install libwebkit2gtk-4.1-0). For an AppImage, check that it's executable (chmod +x). |
| The website shows "Couldn't reach GitHub" | GitHub is unreachable or rate-limited (60 requests/hour per IP). Use the Open GitHub Releases link. See below. |
| Problem | Fix |
|---|---|
| Microphone not allowed | Allow it in your OS privacy settings (macOS: System Settings → Privacy & Security → Microphone). Then use Grant microphone access on the New Meeting screen or in Settings → Audio. |
| "Microphone is clipping (too loud)" | Lower the input volume in your system sound settings (about 50–70%) and turn off automatic mic gain in Zoom/Teams. Distortion can't be removed afterwards. |
| Echo in Mic + Device recordings | Wear headphones, so the mic doesn't pick up your speakers. |
| Device Audio not available on macOS | System audio capture needs macOS 14.2+. Use Speaker mode or Import audio file. |
| The app crashed while recording | Reopen it and click Recover Recording on the banner. |
| "Some audio could not be saved to disk" | Usually the disk is full. Free some space; the meeting keeps what was saved. |
| Problem | Fix |
|---|---|
| "Transcription engine not found" | Release builds include whisper.cpp. If you built from source, install voxtype or a whisper-cli build. Recording works without one. |
| No model in the list | Download one in Settings → Models. Only downloaded models appear. |
| Model download fails | Check your internet connection. Models come from Hugging Face. Set WHISPER_MODEL_DIR if the default folder isn't writable. |
| Transcription failed | The recording is safe. Press Retry, or try another model or a fixed language. |
| Slow transcription | Use a smaller tier (B/C), or turn on GPU acceleration (see Transcription#gpu-acceleration). |
| Wrong language | Choose the language instead of auto-detect, then Re-transcribe. |
| Problem | Fix |
|---|---|
| LLM: Test connection fails | For Ollama, check that it's running (ollama serve) and the model is pulled. Check the base URL, completions path and API key. |
| GitLab: 401 / 403 | The token needs the api scope and access to the project. Check the instance URL for self-hosted GitLab. |
| Calendar: can't find the server | Try Detect from server, Find in Thunderbird or Use SOGo address, or paste the full calendar collection URL. |
| Problem | Fix |
|---|---|
| "Your meetings list could not be loaded." |
Start with a fresh database. The old one is moved to a backup folder, not deleted. Finished meetings are also in Documents/Local Transcribe/. |
| Where are my files? |
~/Documents/Local Transcribe/<title-id>/. Use Settings → App → Local files → Open folder. |
Still stuck? Open an issue in the app repository.