A real-time global dictation utility for macOS (Apple Silicon) that streams audio to Google Cloud Speech-to-Text V2 and types the text into your active window.
- Real-time Dictation: Streams audio to Google Cloud STT V2 (Chirp 2 model).
- Live Typing: Simulates typing with interim results (backspaces and retypes).
- System Tray: Runs in the menu bar with a status indicator.
- Global Hotkeys:
- Toggle Recording:
Cmd + Shift + Space - Cancel Recording:
Escape
- Toggle Recording:
This tool requires portaudio for audio capture.
brew install portaudio- Create a Google Cloud Project.
- Enable the Cloud Speech-to-Text API (specifically V2).
- Set up authentication:
- Create a Service Account.
- Download the JSON key file.
- Set the environment variable:
Note: For the .app bundle, you might need to launch it from a shell that has this variable set, or hardcode/configure it within the app.
export GOOGLE_APPLICATION_CREDENTIALS="/path/to/your/service-account-key.json"
- Clone the repository.
- Run the build script:
./build_app.sh
- Move
Chrisper.appto your Applications folder:mv Chrisper.app /Applications/
- Launch: Open
Chrisper.appfrom your Applications folder. - Permissions:
- The first time you run it, macOS will prompt for Microphone access.
- It will also likely request Accessibility and Input Monitoring permissions. You must grant these in System Settings > Privacy & Security.
- Tip: If it doesn't type, remove "Chrisper" from Accessibility/Input Monitoring and add it back.
- Dictate:
- Look for the icon in the menu bar.
- Press Cmd + Shift + Space to start.
- Speak and watch it type!