Skip to content
 
 

Repository files navigation

MacClaw

Complex Workflows. One simple command.

MacClaw is a 100% native CLI autonomous agent for macOS. It bridges high-level semantic reasoning with pixel-perfect accuracy to automate your desktop and browser without the weight of an Electron app or complex installations.

NPM Version License: MIT

🚀 Quick Start

No heavy installs. Just run via npx:

npx macclaw chat

On your first run, MacClaw will ask for your OpenRouter API Key to power its reasoning engine.

✨ Key Features (v0.3.3)

  • Zero Friction: Run instantly from any terminal. No GUI wrappers or "Install" buttons needed.
  • Intent Reasoning: Leverage world-class vision models like Claude 3.5 Sonnet or Gemini 2.0 Flash via OpenRouter for high-level semantic planning.
  • Grounding Bridge: Powered by UI-TARS 7B, MacClaw accurately translates these model thoughts into physical macOS mouse coordinates.
  • On-Device Voice (TTS): Hear the agent explain its thoughts in real-time with integrated Kokoro-82M (100% local, zero-latency).
  • Robust Bash Engine: Advanced shell execution with nullglob and native afplay integration.
  • Mid-Flight Hotswapping: Use /model or /key commands directly in the prompt to switch engines or credentials without restarting.

📖 Documentation

For full installation guides, grounding architecture deep-dives, and command references, visit our documentation:

👉 macclaw.dev

🛠️ Development

If you'd like to contribute or run the source locally:

  1. Clone the repo:
    git clone https://github.com/heisenbag/macclaw.git
    cd macclaw
  2. Install dependencies:
    pnpm install
  3. Run in dev mode:
    npm run dev

🤝 Contributing

This repository is built for the community. We welcome PRs for new slash commands, better grounding mutations, or UI improvements to the docs site.


Built with ❤️ for macOS power users.

About

The lightweight way to command your macOS GUI and browser through the power of AI.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages