How to Run OpenClaw on Mac M1 (Apple Silicon Guide)

Mac M1 is one of the best machines for running OpenClaw locally. It is fast, uses very little power, and handles agent workloads well.

OpenClaw runs natively on Apple Silicon — no Rosetta, no workarounds. Just install and go. If you want to skip local setup entirely, Ampere.sh gets you running in 60 seconds.

What You Need Before You Start

Requirement	Details
Mac	Any Mac with M1, M2, M3, or M4 chip
macOS	Monterey 12.0 or newer
RAM	8 GB minimum — 16 GB recommended
Node.js	Node 22 LTS or newer (installer handles this automatically)
API key	From Anthropic, OpenAI, Google, or another supported provider
Sleep setting	Set Mac to never sleep so the agent stays online

How to Install OpenClaw on Mac M1

Step 1 — Run the Installer

Open Terminal and run this command. It installs OpenClaw and everything it needs, including Node.js if it is not already on your Mac.

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

The installer detects Apple Silicon automatically. No extra flags or settings needed.

OpenClaw installer running in Mac Terminal — Apple Silicon detected

Step 2 — Run Onboarding

After the install finishes, run the setup command. The --install-daemon flag makes OpenClaw start automatically every time your Mac boots up.

openclaw onboard --install-daemon

The setup will ask you to:

Accept the disclaimer
Choose your AI provider and paste your API key
Confirm the default gateway port (18789)
Connect a messaging app — Telegram, Discord, WhatsApp, or others

OpenClaw onboarding setup — API key and messaging app configuration

Step 3 — Prevent macOS from Sleeping

This is the most important step for 24/7 uptime. macOS will put your Mac to sleep when idle and kill the agent. Turn this off in System Settings.

Open System Settings
Go to Battery → Options
Turn on Prevent automatic sleeping when display is off

macOS Battery settings — Prevent automatic sleeping turned on

Step 4 — Check That It Is Running

Confirm OpenClaw is active with this command:

openclaw gateway status

You can also open the dashboard in your browser to see your agent live:

openclaw dashboard
# Or visit: http://127.0.0.1:18789/

openclaw gateway status showing running on Mac M1

Common Problems and How to Fix Them

openclaw: command not found after install

This happens when your shell does not know where the OpenClaw binary is. On Apple Silicon, the PATH is not always updated automatically after install.

Fix

Run source ~/.zprofile in Terminal, then try the openclaw command again. If it still fails, restart Terminal completely.

source ~/.zprofile

# If still not found, add it manually
export PATH="$(npm prefix -g)/bin:$PATH"

Agent goes offline after screen locks

macOS puts the machine to sleep when idle by default. This kills the OpenClaw gateway even when the daemon is installed.

Fix

Go to System Settings → Battery → Options → turn on Prevent automatic sleeping when display is off. This keeps your agent online when the screen is off.

# Confirm the gateway is running after enabling the setting
openclaw gateway status

Invalid API key error

OpenClaw cannot connect to your AI provider. This usually means the key was entered incorrectly or has been revoked.

Fix

Re-run onboarding to enter your API key again from scratch.

openclaw onboard --install-daemon

Not sure what is broken

Sometimes multiple things go wrong at once — wrong Node version, bad config, PATH issues, gateway not running. Hard to know where to start.

Fix

Run openclaw doctor. It checks your full setup and tells you exactly what needs fixing.

openclaw doctor

If you are not sure what is wrong, run openclaw doctor. It checks everything and tells you exactly what needs to be fixed.

Mac M1 Local Setup vs Ampere.sh

Mac M1 is a solid choice for running OpenClaw locally. Here is how it compares to running it on Ampere.sh.

Feature	Mac M1 (Local)	Ampere.sh
Setup time	10–20 min	60 seconds
Works when Mac sleeps	❌ Goes offline	✅ Always on
Errors to debug	A few possible	None
Maintenance needed	Manual updates	Fully managed
True 24/7 uptime	Only if Mac stays on	✅ Yes
Free to start	API costs only	✅ Free credits

Want to Run OpenClaw on Mac M1 Without Any Local Setup?

Local setup on Mac M1 is straightforward — but your agent is tied to your machine. If your Mac sleeps, restarts, or loses power, the agent goes offline.

Deploy on Ampere.sh instead:

No local install — no Node.js, no config files, no sleep settings
Free credits to start — no credit card required
Runs on cloud servers that never go offline
Your agent stays live even when your Mac is shut down

Frequently Asked Questions

Does OpenClaw work on Mac M1?

Yes. OpenClaw runs natively on Mac M1 with no Rosetta translation needed. The official installer handles Node.js setup automatically for Apple Silicon.

What Node.js version do I need on Mac M1?

Node 22 LTS or newer. Node 24 is recommended. The OpenClaw installer will install the right version automatically if Node is missing.

Can I keep OpenClaw running 24/7 on Mac M1?

Yes. Run openclaw onboard --install-daemon to install as a background service. Also go to System Settings → Battery → Options and enable Prevent automatic sleeping so macOS does not put your Mac to sleep.

Does the M1 chip cause any compatibility issues with OpenClaw?

No. OpenClaw is built on Node.js which fully supports Apple Silicon. You do not need Rosetta or any workarounds.

Where is the OpenClaw config file on Mac?

It is at ~/.openclaw/openclaw.json. You can edit it anytime or re-run openclaw onboard to reconfigure.

What happens to my agent if my Mac M1 restarts?

If you used the --install-daemon flag during setup, the gateway starts automatically on every reboot. For true always-on without depending on your Mac, Ampere.sh runs your agent on cloud infrastructure.

How do I check if OpenClaw is running on Mac M1?

Run openclaw gateway status in Terminal. You can also open the dashboard with openclaw dashboard or visit http://127.0.0.1:18789/ in your browser.