open-claw-mac-mini-setup.md

Beginner’s Guide: Setting Up OpenClaw on a New Mac Mini

OpenClaw is an open-source, locally-running AI agent that lives in your chat apps (Telegram, WhatsApp, Discord, Slack, etc.) and can actually do things on your Mac: send emails, manage calendar, control browser, run shell commands, etc.

Official resources

• Website: https://openclaw.ai/
• Docs: https://docs.openclaw.ai/
• Getting Started: https://docs.openclaw.ai/start/getting-started
• GitHub: https://github.com/openclaw/openclaw

Prerequisites (must have before starting)

Hardware / OS

• New Apple Silicon Mac Mini (M1–M4)
• macOS Ventura 13.0 or later (Sequoia 15+ strongly recommended)
• Stable internet connection
• Administrator account (you’ll be asked for your password)

Software you need

• Terminal app (pre-installed)
• Xcode Command Line Tools (required for native builds)
  Run this first in Terminal:

  xcode-select --install

  Click “Install” when the popup appears.

What the installer will handle automatically

• Homebrew (if you don’t already have it)
• Node.js 22+ (minimum required version)
• OpenClaw CLI (global npm package)

Optional but recommended

• API key from Anthropic (Claude) or OpenAI for best performance
• Ollama for completely local models: https://ollama.com/download

Step-by-Step Setup

1. Run the Official One-Liner
Open Terminal and paste:

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

This installs Homebrew (if missing) → Node.js 22+ → OpenClaw CLI.
It takes 1–3 minutes.

2. Run the Onboarding Wizard

openclaw onboard --install-daemon

The wizard will:

• Ask for your model provider and API key
• Create workspace at ~/.openclaw
• Set up the Gateway (default port 18789)
• Install as a LaunchAgent (runs in background after reboot)
• Install recommended skills
• Guide you through adding a chat channel (Telegram is easiest)

3. Verify It’s Running

openclaw gateway status

Should say “running”.

Open the Control UI (quick test, no channel needed):

openclaw dashboard

or go to http://127.0.0.1:18789 in any browser.

4. Add a Chat Channel
Recommended first channel: Telegram (just needs a bot token).
Do this during onboarding or later with openclaw configure.

5. (Optional) Use Local Models with Ollama

1. Download & install Ollama → https://ollama.com/download
2. Pull a model:

   ollama pull qwen2.5:32b    # or llama3.3:70b, mistral, etc.

3. Set in OpenClaw:

   openclaw config set agents.defaults.model.primary "ollama/qwen2.5:32b"

6. Test It
Send a message to your channel or use the dashboard:
“What’s on my calendar tomorrow?” or “Open Safari and search for flights to Tokyo”.

Troubleshooting

Common Mac Mini issues & fixes

• “command not found: openclaw” → Close & reopen Terminal, or run source ~/.zshrc
• Installer fails → Manually install Homebrew first:

  /bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"

  Then: npm install -g openclaw@latest
• sharp / node-gyp errors →

  SHARP_IGNORE_GLOBAL_LIBVIPS=1 npm install -g openclaw@latest

• Gateway won’t start →

  openclaw gateway install
  openclaw gateway start

  Logs: openclaw gateway --verbose
• macOS permission popups → Grant Full Disk Access, Accessibility, and Screen Recording to “OpenClaw Gateway” in System Settings → Privacy & Security.
• Ollama not found → Make sure ollama serve is running. Test: curl http://localhost:11434/api/tags
• High CPU / memory → Use smaller models (14B or less) or switch to cloud API.
• Doesn’t start after reboot →

  launchctl load ~/Library/LaunchAgents/com.openclaw.gateway.plist

• Reset everything → openclaw onboard --reset

Security note
OpenClaw gets real access to your Mac. Start with one channel and review skills carefully.

You’re done! Your Mac Mini now has a personal AI agent that runs 24/7.

Start simple → one channel + one model → then expand.

Still stuck? Run openclaw doctor or check the wizard guide: https://docs.openclaw.ai/start/wizard