[good]
openclawinstalltutorial

How to Install OpenClaw: Step-by-Step Setup Guide (2026)

February 23, 2026 · 6 min read

Installing OpenClaw takes about five minutes. You need Node.js 22+ and a terminal. The installer script handles everything else — it detects your Node version, installs the CLI, and launches the onboarding wizard.

Requirements

  • Node.js 22 or newer. Check with node --version. If you need to install Node, the installer script will offer to do it for you.
  • macOS, Linux, or Windows (WSL2). Native Windows works, but Microsoft's WSL2 gives the most reliable experience on Windows machines.
  • An API key from at least one AI provider: Anthropic, OpenAI, or Google. If you want to run local models, you can use Ollama instead — no key needed.

Step 1: Run the installer

On macOS or Linux, open Terminal and run:

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

On Windows, open PowerShell and run:

iwr -useb https://openclaw.ai/install.ps1 | iex

The script downloads and installs the OpenClaw CLI globally. If Node 22+ is not detected, it will offer to install it. You do not need to install anything manually first.

Step 2: Run the onboarding wizard

After the installer finishes, launch the wizard to configure your API keys, gateway settings, and optionally connect a channel:

openclaw onboard --install-daemon

The --install-daemon flag registers the gateway as a background service so it starts automatically when your machine boots. This is the recommended setup for anything beyond a quick test.

The wizard will ask for:

  1. An AI provider API key (Anthropic, OpenAI, or Google Gemini)
  2. A default model
  3. Whether to install the gateway daemon
  4. Optionally, a messaging channel (Telegram, Discord, etc.)

Step 3: Check the gateway is running

openclaw gateway status

You should see the gateway reported as running. If it shows stopped, start it manually:

openclaw gateway start

Step 4: Open the Control UI

openclaw dashboard

This opens the browser-based Control UI at http://127.0.0.1:18789. If it loads, your gateway is fully operational. You can chat directly from the dashboard without setting up any messaging channel — useful for testing your setup.

Step 5: Add your API key (if you skipped it in the wizard)

To set a Claude (Anthropic) key:

openclaw config set auth.providers.anthropic.apiKey "sk-ant-..."

For OpenAI:

openclaw config set auth.providers.openai.apiKey "sk-..."

For Google Gemini:

openclaw config set auth.providers.google.apiKey "AIza..."

Restart the gateway after any config change:

openclaw gateway restart

Common errors and fixes

node: command not found or Node version too old
Re-run the installer — it will detect the missing or outdated Node version and offer to install the correct one. Alternatively, install Node 22+ manually from nodejs.org.
Gateway fails to start — port in use
The default gateway port is 18789. If another process is using it, set a different port: openclaw config set gateway.port 18790, then restart.
API key not working after config set
Make sure you restarted the gateway after setting the key. The gateway reads config at startup — changes don't apply to a running instance.
Windows: permission errors with WSL2
Run WSL2 as administrator for the initial install. After installation completes, you can use a normal terminal for everything else.

Next steps

Comparing AI models? See our LLM comparisons →