How to Install OpenClaw on Mac?

Installing OpenClaw on macOS is straightforward — it works natively on both Intel and Apple Silicon Macs. This includes MacBook, iMac, Mac Mini, and Mac Pro.

First, make sure you have Node.js 22 or newer. If you don't, the installer will set it up via nvm automatically. Open Terminal and run the install script. It downloads OpenClaw, installs dependencies, and launches the onboarding wizard.

The onboarding wizard walks you through three steps: choosing your AI model (Claude, GPT, DeepSeek, or local via Ollama), entering the API key, and connecting your first messenger (usually Telegram).

Mac Mini is especially popular as a dedicated OpenClaw server. The M-series chips are power-efficient and capable enough to run local models via Ollama alongside the agent. A Mac Mini M2 with 16GB RAM can comfortably run both OpenClaw and a 7B parameter model 24/7 with minimal power consumption.

For 24/7 operation on Mac, set up a launchd service so OpenClaw starts automatically on boot and restarts on crashes. The CLI includes a helper command for this. Also disable sleep in System Settings to keep the agent running continuously.

If you plan to use Ollama for local models, install it separately from ollama.com before running OpenClaw onboard — the wizard will detect it automatically.

curl -fsSL https://openclaw.sh/install | bash

openclaw onboard

# Enable auto-start on macOS
openclaw service install