How to Install OpenClaw

Step 1: Check requirements first

OpenClaw depends on modern Node features, so Node.js 22 or later is required. On Windows, the supported path is WSL2. Installing on older Node versions is the most common cause of failures.

Step 2: Use the official install script

The recommended method installs the CLI and prepares the runtime:

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

Step 3: Complete onboarding

The wizard walks you through pairing a messaging channel, selecting a model provider, and enabling skills. This is where you set initial permissions and confirm what OpenClaw can access.

Helpful commands after install

Use openclaw doctor to validate dependencies, openclaw status or openclaw health to check the gateway, and openclaw update to stay current.

Verify the install before production use

After onboarding, run a small test task that uses a low-risk skill, such as creating a file in a temporary folder or summarizing a local text file. This confirms that your channel, model, and skills are correctly wired. If you are deploying on a server, verify that only required ports are open and that admin pairing is enforced.

Common installation pitfalls

Most failures come from an outdated Node version, missing build tools, or running the script under the wrong user account. Another frequent issue is partial installs caused by network errors during package download. If the CLI behaves inconsistently, re-run the install script and check logs with openclaw doctor.

Security reminders

Do not expose admin endpoints to the public internet, and avoid installing untrusted skills. If you run OpenClaw on a server, lock down ports and use least-privilege access.

Official references

For exact instructions, see the official install and wizard pages: Install and Onboard.