FAQ

Frequently asked questions about Yap

General

What is Yap?

Yap is a desktop speech-to-text application. Press a hotkey, speak, and your words are transcribed and pasted into any application.

Is it free?

Yes! Yap is free and open source under the MIT license.

What platforms are supported?

macOS: Apple Silicon (native) and Intel (via Rosetta)
Windows: Windows 10/11 (x64)
Linux: AppImage and DEB packages

Does it work offline?

Yes, when using local transcription with whisper.cpp. Models are downloaded once and work entirely offline.

Privacy

Is my audio sent anywhere?

Local mode: No. Everything stays on your device.

Cloud mode: Audio is sent to OpenAI’s servers for transcription. OpenAI does not use API data for training per their data usage policy.

Where is my data stored?

All data is stored locally in:

macOS: ~/Library/Application Support/yap/
Windows: %APPDATA%\yap\

Nothing is synced to the cloud unless you use OpenAI transcription.

Can I delete my recordings?

Yes. Delete individual recordings from History, or delete all with Clear History. You can also delete files directly from the data folder.

Transcription

Which is better: local or cloud?

Aspect	Local	Cloud
Privacy	Better	Good (sent to OpenAI)
Accuracy	Good	Better
Speed	Depends on hardware	Fast
Cost	Free	~$0.006/minute
Offline	Yes	No

Recommendation: Start with local. Try cloud if you need better accuracy.

Which model should I use?

For most users: base.en

Quick notes: tiny.en
Daily use: base.en
Important content: small.en or medium.en
Maximum accuracy: large-v3

Why is transcription slow?

Local transcription speed depends on:

Model size (larger = slower)
Your CPU (Apple Silicon is fastest)
Audio length

Try a smaller model or cloud transcription for faster results.

Can it transcribe languages other than English?

Yes! Use multilingual models (without .en suffix):

tiny, base, small, medium, large-v3

For English-only, use .en models for better accuracy.

Can it transcribe from audio files?

Currently no. Yap only transcribes live recordings. File transcription may be added in the future.

Features

Why doesn’t auto-paste work in some apps?

Some applications don’t accept simulated paste commands. This is a limitation of how those apps handle input. The transcribed text is still copied to your clipboard — you can paste manually with Cmd+V.

Can I customize the hotkey?

Yes! Go to Settings → Recording Hotkey. Options include:

Right Option
Left Option
Fn
Double-tap Right Option

Can I use a custom hotkey combination?

Not currently. Only the preset options are available. Custom combinations may be added in a future version.

Does it support voice commands?

No. Yap transcribes speech to text only. It doesn’t interpret commands or control your computer.

Can I train it on my voice?

No. Whisper models are pre-trained and not customizable. However, speaking clearly and using larger models improves accuracy.

Technical

What is whisper.cpp?

whisper.cpp is a C++ implementation of OpenAI’s Whisper speech recognition model. It runs on your CPU without needing a GPU.

What’s the difference between whisper.cpp and OpenAI?

Both use the same Whisper model, but:

whisper.cpp: Runs locally on your device
OpenAI API: Runs on OpenAI’s servers

OpenAI may use a larger model variant and has optimized inference.

Does it use my GPU?

No. whisper.cpp uses CPU only. This means it works on all hardware but is slower than GPU-accelerated alternatives.

How much does OpenAI cost?

OpenAI Whisper API costs approximately $0.006 per minute of audio. For typical use, expect:

Light use: $1-2/month
Heavy use: $5-10/month

Why does it need Accessibility permissions?

macOS Accessibility permissions enable:

Global hotkeys — Detect key presses when the app isn’t focused
Auto-paste — Simulate Cmd+V to paste text

Without these, you’d need to manually start recording from the app window.

Troubleshooting

Contributing

How can I contribute?

See the Contributing Guide. We welcome:

Bug reports
Feature requests
Code contributions
Documentation improvements

Where can I report bugs?

Open an issue on GitHub: github.com/ApplauseLab/yap/issues

Is the source code available?

Yes! Yap is open source: github.com/ApplauseLab/yap