Voice typing from your terminal.

A simple CLI tool that gives you fast voice typing in every Mac app. Instant speech-to-text from your terminal to your cursor with one hotkey - works with Claude Code, Cursor, and any macOS application.

It just works:

npm install -g speech2type && s2t start

• ⚡️ Real-time transcription: Your words appear instantly as you speak
• 💻 Works everywhere: Claude Code, Cursor, Slack, Chrome, Lovable... any Mac app with text input
• 🎯 Inline-typing: Text is inserted directly at the cursor’s position (no clipboard involved).
• 💸 Completely free: Open source with free Deepgram API tier
• 🌍 40+ languages: English, Spanish, French, German, Japanese, Chinese, and many more

• Developers: Vibe coding in any environment (Claude Code CLI, Cursor IDE, etc.)
• Creators: Dictating in any text editor
• Productivity: Quick voice input in Slack, email, and more

npm install -g speech2type

# Start the application
# s2t or s2t start
s2t start

# Follow the setup wizard to:
# 1. Get your FREE Deepgram API key from https://deepgram.com
# 2. Configure your language preference (default: English)
# 3. Set up your hotkey (default: ⌘;)
# 4. Grant microphone and accessibility permissions

1. Start the app: Run s2t start in your terminal
2. Position your cursor: Click in where you want text to appear
3. Press your hotkey: Default is ⌘; to start listening
4. Speak: Your words will appear in real-time at your cursor
5. Press hotkey again: Stop listening

That's it! You now have voice typing in every Mac app.

• macOS 13+ with Apple Silicon
• Node.js 18+
• Deepgram API key (free tier available, no credit card required)
• Xcode Command Line Tools (only for development)

Speech2Type requires two permissions that will be requested on first run:

1. Microphone access: To capture your voice
2. Accessibility access: To listen global hotkeys
3. Automation access - System Events: To inject text into other applications

Grant these permissions in System Settings → Privacy & Security for your terminal app from which you run Speech2Type!

s2t config

# Change hotkey combination (default: ⌘;)
s2t config --hotkey

# Select language for speech recognition (default: English)
s2t config --language

# Update your Deepgram API key (or set the DEEPGRAM_API_KEY environment variable)
s2t config --deepgram-api-key

Speech2Type supports 40+ languages including:

• Bulgarian: bg,
• Catalan: ca,
• Chinese (Mandarin, Simplified):zh, zh-CN,zh-Hans,
• Chinese (Mandarin, Traditional):zh-TW,zh-Hant,
• Chinese (Cantonese, Traditional): zh-HK,
• Czech: cs,
• Danish: da, da-DK,
• Dutch: nl,
• English: en, en-US, en-AU, en-GB, en-NZ, en-IN,
• Estonian: et,
• Finnish: fi,
• Flemish: nl-BE,
• French: fr, fr-CA,
• German: de,
• German (Switzerland): de-CH,
• Greek: el,
• Hindi: hi,
• Hungarian: hu,
• Indonesian: id,
• Italian: it,
• Japanese: ja,
• Korean: ko, ko-KR,
• Latvian: lv,
• Lithuanian: lt,
• Malay: ms,
• Norwegian: no,
• Polish: pl,
• Portuguese: pt, pt-BR, pt-PT,
• Romanian: ro,
• Russian: ru,
• Slovak: sk,
• Spanish: es, es-419,
• Swedish: sv, sv-SE,
• Thai: th, th-TH,
• Turkish: tr,
• Ukrainian: uk,
• Vietnamese: vi

See supported languages for the up-to-date full list.

# Clone the repository
git clone https://github.com/gergomiklos/speech2type.git
cd speech2type

# Install dependencies
npm install

# Build Swift components
npm run build

# Run in development mode
npm run dev

speech2type/
├── bin/                   # Executable scripts
│   └── speech2type        # Main CLI entry point
├── src/                   # JavaScript source code
│   ├── index.js           # Main application logic
│   ├── config.js          # Configuration management
│   └── services/          # Core services
│       ├── hotkey.js      # Global hotkey management
│       ├── transcriber.js # Deepgram integration
│       └── typer.js       # Text injection via osascript
|       └── permission.js  # macOS permission handling
├── data/                  # Language data
│   └── languages.json     # Supported languages
└── swift/                 # Native Swift components
    ├── hotkey-manager.swift     # Global hotkey capture
    ├── mic-recorder.swift       # Audio recording
    └── permission-checker.swift # System permissions

1. "Permission denied" errors

• Grant microphone access to your terminal app in System Settings → Privacy & Security → Microphone

• Grant accessibility access to your terminal app in System Settings → Privacy & Security → Accessibility

  Important: the terminal app must be the one that is running Speech2Type. Common terminal apps: Terminal (built-in macOS terminal), Cursor (integrated in-app terminals), Warp, VS Code...

2. "Command not found: s2t"

• Reinstall globally: npm install -g speech2type
• Check your PATH includes npm global binaries

3. "Hotkey not working"

• Check for conflicts with other applications
• Reconfigure hotkey with: s2t config --hotkey
• Ensure permissions are granted

4. "Text not appearing"

• Ensure your deepgram api key is correct
• Check your microphone is working
• Check you have internet connection
• Ensure permissions are granted

5. "Text not appearing in secure fields"

• This is by design - secure input fields (passwords) don't accept simulated typing
• Speech2Type works in regular text fields only

6. "How to get a Deepgram API key?"

• Go to https://deepgram.com/

• Sign up for a free account

• Go to https://console.deepgram.com/

• Click on "API Keys"

• Click on "Create API Key"

• Copy the API key

• Run s2t config --deepgram-api-key and paste the API key

  The free tier includes more than 50 hours of credits. No credit card required.

FAQ: https://speech2type.com/faq

Run with debug output for troubleshooting:

DEBUG=1 s2t start

• Documentation: This README and inline help (s2t --help)
• Issues: GitHub Issues
• Discussions: GitHub Discussions

We welcome contributions! Here's how to get started:

1. Fork the repository
2. Create a feature branch: git checkout -b feature/amazing-feature
3. Make your changes: Follow the existing code style
4. Test thoroughly: Ensure your changes work on macOS
5. Submit a pull request: Describe your changes clearly

• Code style: Follow existing JavaScript and Swift patterns
• Testing: Test on multiple macOS versions when possible
• Documentation: Update README for new features
• Permissions: Be mindful of security and privacy implications

Deepgram provides real-time streaming, higher accuracy, wider language support with reasonable low cost and developer-friendly APIs, allowing Speech2Type to remain native and lightweight (Speech2Type is not affiliated with Deepgram).

This project is licensed under the MIT License - see the LICENSE file for details.

If Speech2Type helps your workflow, consider:

• Starring the repository
• Reporting bugs and suggesting features
• Improving documentation
• Contributing code
• Sharing with others who might benefit

Speech2Type - If you can type there, you can speak there.

Official website: https://speech2type.com

Built with ❤️