FAQ

Common questions about Clanker Yap.

General

What is Clanker Yap?

Clanker Yap is a local voice-to-text transcription application. It converts your spoken words into text that is automatically pasted into any application. All processing happens on your machine using the Whisper machine learning model.

How does it work?

1. Hold a configurable hotkey to start recording
2. Speak into your microphone — a floating overlay pill appears showing FFT audio bars
3. Release the hotkey to stop recording — overlay shows processing state
4. Audio is transcribed locally using Whisper
5. Text is automatically pasted into your active application
6. Overlay disappears after transcription completes

Is my audio sent to the cloud?

No. All processing happens locally on your machine. Your voice recordings never leave your device.

What languages are supported?

• English models (.en suffix) - Optimized for English only
• Multilingual models - Support 99+ languages

Choose a model based on your language needs.

What's the default hotkey?

The default hotkey is CmdOrCtrl+Shift+V, which resolves to Ctrl+Shift+V on Linux.

Can I use a different hotkey?

Yes. Open the app, click the Hotkey toolbar button, and follow the instructions to set a new shortcut.

Technical

What models are supported?

GGML-format Whisper models from ggml-org/whisper.cpp.

Do I need a GPU?

No. Clanker Yap runs entirely on CPU. No GPU is required.

What audio formats are supported?

• Input: Any format your microphone supports
• Internal: 16kHz 16-bit mono (required by Whisper)
• The app automatically resamples audio to the correct format.

What does the recording overlay show?

A floating pill appears at the bottom of the screen while recording:

• 7 FFT frequency bars — react to your voice in real time (low bars = bass, high bars = treble)
• Recording state — green/cyan gradient bars with a pulsing dot
• Processing state — amber pulsing bars when transcription is running

The overlay is always-on-top, click-through, and transparent. On Linux it uses GTK Layer Shell when that optional feature is enabled, with an always-on-top fallback.

How much memory does it use?

Approximate memory usage:

• Model (base): ~150 MB
• Audio buffer: ~1 MB per recording
• App overhead: ~50 MB

Can I use a different microphone?

Yes. Go to Mic in the toolbar to select a specific audio device, or use the system default.

Usage

How do I use it?

1. Ensure a Whisper model is downloaded (see Getting Started)
2. Start the app
3. Hold the hotkey and speak
4. Release when done
5. Text is automatically pasted (Ctrl+V simulated via persistent input controller)

Can I use it in a terminal?

Yes. Set the paste mode to "Terminal" in settings for better terminal compatibility.

How do I disable automatic paste?

Go to Settings → Paste → Edit and toggle Auto-paste off. The transcription will still be copied to your clipboard — press Ctrl+V to paste manually.

On KDE/Wayland, disabling auto-paste avoids the "Remote Control" permission prompt entirely.

How do I review past transcriptions?

Click History in the toolbar to see recent transcriptions.

Can I copy text without pasting?

Yes. The transcribed text is also copied to your clipboard. You can paste manually with Ctrl+V.

What if the app isn't responding?

1. Wait for current transcription to complete
2. If still unresponsive, close and restart the app
3. Check logs with RUST_LOG=trace npm run tauri:dev

Configuration

Where is settings stored?

Settings are stored in a SQLite database:

• Linux: ~/.local/share/voice-transcribe/voice-transcribe.db

Can I use a custom model path?

Yes. Click Model in the toolbar and enter a custom path.

What's the difference between paste modes?

Auto-paste (a separate toggle in Paste settings) controls whether Clanker Yap automatically simulates Ctrl+V after copying text to the clipboard.

How do I reset settings?

Delete the database file and restart the app:

rm ~/.local/share/voice-transcribe/voice-transcribe.db

Troubleshooting

Hotkey doesn't work

1. Ensure no other app is using the same shortcut
2. Try a different hotkey
3. Check app permissions
4. See Troubleshooting

Microphone not detected

1. Check microphone is connected
2. Test microphone in system settings
3. Try selecting specific device in app
4. See Troubleshooting

Model won't load

1. Verify model file exists at configured path
2. Download model if missing
3. Check file permissions
4. See Troubleshooting

Text not pasting

1. Try "Terminal" paste mode
2. Check target app accepts paste
3. Manually paste with Ctrl+V
4. See Troubleshooting

Development

Is this open source?

Yes. Clanker Yap is open source. See LICENSE.

Can I contribute?

Yes! See CONTRIBUTING.md for guidelines.

What technologies are used?

How is the project structured?

See Architecture for detailed project structure.

Privacy

Does the app collect data?

No. Clanker Yap does not:

• Send audio to any server
• Collect usage statistics
• Track your transcriptions (except locally)
• Report errors externally

Is my transcription history private?

Transcriptions are stored locally in the SQLite database. No data is sent externally.

Support

Where can I get help?

1. Read the docs
2. Check Troubleshooting
3. Search existing issues
4. Create a new issue with details

How do I report a bug?

1. Check existing issues
2. Run with RUST_LOG=trace to capture logs
3. Create issue with:
   • Platform and OS version
   • Steps to reproduce
   • Expected vs actual behavior
   • Log output

How do I request a feature?

Create a feature request issue with:

• Description of the feature
• Use case / why you need it
• Any implementation suggestions

Future

Will there be a macOS or Windows build?

There are no macOS or Windows release targets. Clanker Yap ships only as a Linux x86_64 AppImage.

Can I use a larger model?

Yes. Download any GGML Whisper model and set the path in settings.

Will GPU support be added?

Currently CPU-only. GPU acceleration could be added but is not planned for v0.1.0.

Can I use WhisperX or other backends?

No. Clanker Yap uses whisper-rs specifically. Other backends would require significant changes.