Yet another dictation app for macOS.
yada records audio in memory, transcribes with OpenAI, rewrites for punctuation/casing, and inserts text at the current cursor. It is designed for direct download (not App Store), minimal UI, and clean code.
Demo (click to watch):
yada-demo-5.mp4
- Download the latest
yada-*-mac.zipfrom GitHub Releases. - Double-click the zip to unzip it.
- Drag
yada.appinto your Applications folder. - First launch: right-click
yada.app→ Open → Open. - If macOS blocks it, go to System Settings → Privacy & Security and click Open Anyway.
- Open yada from Applications.
- Enter your OpenAI API key and click Save.
- Grant Microphone and Accessibility permissions when prompted.
- Press the global hotkey to start/stop dictation (default: Cmd+grave).
- Global hotkey (configurable) with two modes:
- Toggle mode: press to start, press again to stop (default).
- Hold mode: hold to record, release to stop.
- Mic selection (sets system default input device).
- In-memory audio capture (no audio/text saved to disk).
- OpenAI transcription + rewrite pipeline.
- Configurable rewrite prompt for customizing GPT instructions.
- Accessibility-based insertion with clipboard fallback.
- Transcription model:
gpt-4o-transcribe - Rewrite model:
gpt-5-mini - Hotkey: Cmd+grave (backtick key) with at least one modifier required
- Hotkey mode: Toggle (press to start/stop)
- Rewrite prompt: "Rewrite the text with correct punctuation and capitalization. Preserve meaning. Return plain text only."
See docs/DEVELOPMENT.md for developer intro, build, and run instructions.
Windows scaffold info is in docs/windows-contributing.md.