Privacy-first dictation for macOS
VoiceType is an open-source, privacy-first dictation tool for macOS that converts speech to text using local AI models. Unlike cloud-based solutions, VoiceType processes everything on your device, ensuring your voice never leaves your computer.
- 🔒 100% Privacy: All processing happens on-device. No cloud, no data collection, no internet required
- 🚀 Fast & Accurate: Real-time transcription with <5 second latency using OpenAI Whisper models
- 🌍 30+ Languages: Built-in support for multiple languages with auto-detection
- ⌨️ Universal Compatibility: Works with any macOS application that accepts text input
- 🎛️ Flexible Models: Choose between speed and accuracy with multiple model sizes
- 🔌 Extensible: Plugin system for custom audio processors and text injectors
- 📖 Open Source: MIT licensed, community-driven development
Download the latest release from the Releases page.
- Open VoiceType - Look for the microphone icon in your menu bar
- Grant Permissions - Allow microphone access when prompted
- Choose Your Model - Select Fast (default) for quick results or Accurate for better quality
- Set Your Hotkey - Default is
Ctrl+Shift+V - Start Dictating - Press your hotkey in any app and start speaking!
- macOS 12.0 (Monterey) or later
- 8GB RAM minimum (16GB recommended for larger models)
- Apple Silicon (M1/M2/M3) or Intel processor
- ~200MB disk space (plus model downloads)
- Position your cursor where you want to insert text
- Press your hotkey (default:
Ctrl+Shift+V) - Speak clearly for up to 5 seconds
- Watch your words appear - VoiceType automatically inserts the text
- Speak naturally at a normal pace
- Minimize background noise for best results
- Use the Accurate model for technical terms
- Customize your hotkey in Settings
- Xcode 15.0 or later
- macOS 13.0+ (for development)
- Apple Developer account (for code signing)
# Clone the repository
git clone https://github.com/yourusername/voicetype.git
cd voicetype/VoiceType
# Setup development environment
./Scripts/setup.sh
# Build the app
./Scripts/build.sh
# Run tests
./Scripts/test.sh
# Create release build
./Scripts/release.sh# Open in Xcode
open VoiceType.xcodeproj
# Or use Swift Package Manager
swift build
swift testVoiceType can be customized through its settings panel or by editing the configuration file:
~/Library/Application Support/VoiceType/config.json
- Hotkey: Customize your recording trigger
- Model Selection: Choose between Tiny (fast), Base (balanced), or Small (accurate)
- Language: Select from 30+ languages or use auto-detection
- Audio Device: Choose your preferred microphone
We welcome contributions! Please see our Contributing Guide for details.
- 🔌 App-specific text injectors - Add support for more applications
- 🎤 Audio preprocessors - Improve noise reduction and audio quality
- 🌐 Translations - Help translate the UI to more languages
- 📚 Documentation - Improve guides and tutorials
- 🐛 Bug fixes - Help us squash bugs
VoiceType uses a modular, protocol-first architecture:
┌─────────────┐ ┌──────────────┐ ┌──────────────┐
│ Menu Bar │────▶│ Coordinator │────▶│Audio Processor│
└─────────────┘ └──────────────┘ └──────────────┘
│ │
▼ ▼
┌──────────────┐ ┌──────────────┐
│ Transcriber │────▶│Text Injector │
└──────────────┘ └──────────────┘
See our Architecture Guide for details.
VoiceType doesn't appear in menu bar
- Check if the app is running in Activity Monitor
- Try launching from Applications folder
Hotkey doesn't work
- Grant Input Monitoring permission in System Settings → Privacy & Security
- Check for conflicts with other apps
No text appears after speaking
- Verify microphone permission is granted
- Check audio input levels in Settings
- Try the clipboard fallback mode
See our Troubleshooting Guide for more solutions.
VoiceType is released under the MIT License. See LICENSE file for details.
- OpenAI Whisper for the amazing speech recognition models
- Apple CoreML for on-device inference
- The Swift and macOS developer communities