# OpenClaw on Mac M2: Simple Setup Guide for Apple Silicon > Mac M2 handles AI workloads better than almost anything else at its price. Here is how to put that power to use. **Published:** Mar 24, 2026 | **Author:** Michael Park | **Read time:** 10 min read Run OpenClaw on Mac M2 with simple steps. Install in minutes, connect your messaging app, and keep your AI agent running all day on Apple Silicon. --- The Mac M2 chip is fast, power-efficient, and built for sustained workloads — which makes it one of the best personal machines for running an AI agent like [OpenClaw](/blog/what-is-openclaw). This guide takes you from a fresh Mac M2 to a fully working agent. No prior experience needed. If you prefer cloud hosting over local setup, [ Ampere.sh ](https://www.ampere.sh) lets you skip all of this and get started right away. The Mac M2 chip is fast, power-efficient, and built for sustained workloads — which makes it one of the best personal machines for running an AI agent like [OpenClaw](/blog/what-is-openclaw). This guide takes you from a fresh Mac M2 to a fully working agent. No prior experience needed. If you prefer cloud hosting over local setup, [ Ampere.sh ](https://www.ampere.sh) lets you skip all of this and get started right away. ## Why Mac M2 is Great for OpenClaw Most personal computers were not built for running background AI services. Mac M2 is different. Here is what makes it a solid choice: ## Mac M2 Requirements for OpenClaw Check these before starting so the install goes smoothly: - **macOS version** — Ventura 13.0 or newer. Monterey 12.0 also works but Ventura is recommended. - **RAM** — 8 GB is the minimum. 16 GB gives you room for multiple agents or heavier models. - **Storage** — OpenClaw uses under 200 MB. Make sure you have at least 1 GB free. - **API key** — You need one from Anthropic, OpenAI, Google, or another supported provider before you start. - **Terminal** — Open the built-in Terminal app (Applications → Utilities → Terminal) or use iTerm2. ## Installing OpenClaw on Mac M2 ### Step 1 — Run the Installer Open Terminal and paste this command. It automatically installs OpenClaw along with the right version of Node.js for your M2 chip. The script detects Apple Silicon and installs the native arm64 build. No Rosetta, no compatibility mode. ### Step 2 — Run the Setup Command Once the install finishes, run this command in Terminal. Do not remove the `--install-daemon` part — it makes OpenClaw start automatically every time your Mac boots up. ### Step 3 — Accept the Disclaimer The setup wizard will show a short disclaimer first. Read it and press **Enter** to continue. ### Step 4 — Choose Your AI Provider You will see a list of AI providers. Use the arrow keys to pick one — for example Anthropic, OpenAI, or Google — then press Enter. ### Step 5 — Paste Your API Key Copy your API key from your provider dashboard and paste it when asked. The key will not be visible as you type — that is normal. ### Step 6 — Set the Gateway Port Press **Enter** to keep the default port `18789`. You do not need to change this unless another app is already using that port. ### Step 7 — Connect a Messaging App Choose which app you want to use to chat with your agent — Telegram, Discord, WhatsApp, or others. Follow the steps shown on screen to link your account. ### Step 8 — Verify the Agent is Live Run this to confirm everything is working: You should see the gateway running with a PID and port number. Open the web dashboard to manage your agent from a browser: ## Keeping Your Agent Online 24/7 The daemon flag keeps OpenClaw running in the background, but macOS can still put your Mac to sleep and take the agent offline. Fix this in one place: - Open **System Settings** - Go to **Battery → Options** - Turn on **Prevent automatic sleeping when display is off** On a MacBook M2, keep it plugged in when you want the agent running overnight. On a [Mac Mini M2](/blog/openclaw-on-mac-mini), this setting is all you need — the machine is already always on. ## Fixes for Common Issues ## Mac M2 Local Setup vs Ampere.sh | Feature | Mac M2 (Local) | Ampere.sh | | --- | --- | --- | | Time to get started | 15–20 minutes | 60 seconds | | Agent offline when Mac sleeps | ❌ Yes | ✅ Never | | Setup required | Terminal + config | None | | Agent uptime | Depends on your Mac | ✅ 24/7 guaranteed | | Maintenance | You manage updates | Fully managed | | Cost to start | API key only | ✅ Free credits | Running OpenClaw on your own Mac M2 gives you full control. Running it on Ampere.sh means you never have to think about uptime again. Here is the tradeoff: ## Your Mac M2 is Powerful. Your Agent Should Not Depend on It. Running OpenClaw locally on Mac M2 works well — until your Mac restarts, the lid closes, or the power goes out. When that happens, your agent goes silent. [ Ampere.sh ](https://www.ampere.sh/setup) keeps your agent running on dedicated cloud servers — completely separate from your Mac. No config files, no sleep settings, no Terminal required: - Your agent answers messages even when your Mac is shut down - Nothing to install or maintain — updates happen automatically - Start for free, no credit card required - Works alongside your local setup or as a full replacement ## Run OpenClaw Without Local Maintenance If you want OpenClaw always online without sleep settings, restarts, or local maintenance, run it on Ampere.sh and keep your agent available 24/7. [Start your setup](https://www.ampere.sh/setup) ## Frequently Asked Questions ### Does OpenClaw run natively on Mac M2? Yes. OpenClaw is built on Node.js which has full native support for Apple Silicon. It runs without Rosetta and takes full advantage of the M2 chip. ### Is Mac M2 better than M1 for running OpenClaw? Both work great. M2 has faster memory bandwidth and better efficiency, which means lower latency during conversations and slightly less battery usage. For most users the difference is not noticeable in daily use. ### Does OpenClaw work on M2 Pro and M2 Max? Yes. OpenClaw works on all M2 variants — M2, M2 Pro, and M2 Max. More RAM on Pro and Max means you can run more agents or larger models at the same time. ### Will my agent stop working if I close the MacBook lid? Yes, unless you configure Power Nap or connect to a charger and enable Prevent automatic sleeping in System Settings. On a Mac Mini with M2, this is easier since the machine is always plugged in. ### How do I update OpenClaw on Mac M2? Run npm install -g openclaw in Terminal. The update process is the same on all Mac models. ### Can I run multiple OpenClaw agents on M2? Yes. M2 with 16 GB or more RAM handles multiple agents well. Each agent uses around 150 to 300 MB at idle. Keep an eye on memory if you plan to run more than three at the same time. ### What is the config file location on Mac M2? The config file is at ~/.openclaw/openclaw.json. You can edit it directly or re-run openclaw onboard to change settings through the setup wizard.