Deploying OpenClaw AI Agents
Environment and Gateway Setup
First, Your Environment
Before you can deploy your AI agent, you need to set up the right environment. OpenClaw runs on Node.js, and it specifically requires version 22 or newer. You can check which version you have installed by opening your terminal and running this command:
node -v
If you see a version number like v22.0.0 or higher, you're ready to proceed. If not, you'll need to update your Node.js installation first. Once that's sorted, you can install the OpenClaw Command Line Interface (CLI) globally on your system. This tool is your main entry point for managing your agent. Use either npm or pnpm to install it:
# Using npm
npm install -g openclaw
# Or using pnpm
pnpm add -g openclaw
The Onboarding Wizard
The simplest way to get your agent running is with the built-in onboarding wizard. This interactive setup process will configure everything for you, from connecting your AI model to setting up the background service.
Preferred setup: run the onboarding wizard (openclaw onboard) in your terminal.
To start the wizard, just run the following command:
openclaw onboard
The wizard will first ask you to choose a Large Language Model (LLM) provider. OpenClaw supports major platforms like OpenAI, Anthropic, and Google Gemini. You will need to provide an API key for the service you choose. This key allows your OpenClaw agent to communicate with the AI model.
Running the Gateway
One of the final steps in the onboarding wizard is setting up the for persistence. This is a critical background process that ensures your agent runs 24/7, even if you close your terminal or restart your computer. The wizard automatically configures this for you using the native service manager for your operating system: launchd on macOS or systemd on Linux.
When prompted, you should almost always choose to install the daemon. This turns OpenClaw from a command you run into a persistent agent that's always on.
All your settings are saved in a configuration file named openclaw.json in your home directory. While you can edit this file manually, the onboarding wizard is the recommended way to handle the initial setup. After the wizard completes, the Gateway will be running.
You can now access the Web Control UI to monitor your agent. Open a browser and navigate to http://localhost:18789. This interface provides a real-time view of your agent's activity, including its internal logic, the tools it's using, and any logs it generates. It's an essential tool for debugging and understanding what your agent is doing at any given moment.
What is the minimum required version of Node.js to run OpenClaw?
Which command initiates the interactive setup wizard for a new OpenClaw agent?
With the environment set up and the Gateway running, you're ready to start connecting your agent to messaging platforms and giving it tasks.