A Brief Guide to Getting Started with Openclaw

OpenClaw isn't just another AI chatbot. It's a self-hosted gateway that transforms your personal machine into a private, multi-channel command center for AI agents. Think of it as giving your favorite coding assistant—like Pi or yourself—direct access to your messages on WhatsApp, Telegram, iMessage, and more, all while keeping your data under your control.

The best part? You don't need a server or to give an external app access to your private conversations. Everything runs on your own hardware, on your terms.

This guide will get you from zero to your first live chat in under five minutes.

Prerequisites: Node.js

Before anything else, you need Node.js version 22 or newer. This is the engine that powers OpenClaw.

To check your version, open your terminal and run:

node --version

If you don't have Node installed, or you're on an older version, head to nodejs.org to download the latest LTS release.

Step 1: Install the OpenClaw CLI

The fastest way to install OpenClaw is through its dedicated script. Copy and run this command in your terminal:

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

This one-liner will download and set up the OpenClaw command-line interface (CLI) on macOS or Linux. (Windows users can use PowerShell with a different command, but for simplicity, we'll focus on the most common setup.)

Step 2: Run the Onboarding Wizard

With the CLI installed, activate the interactive setup wizard. This is where the magic happens—OpenClaw will walk you through authentication, core settings, and your first channel connection.

openclaw onboard --install-daemon

The --install-daemon flag tells the installer to set up OpenClaw as a background service on your machine. This means it will start automatically when your computer boots, acting like a silent AI butler that's always ready.

During this process, you'll be prompted to:

Choose a secure password for the web dashboard.
Set up your first messaging channel (like WhatsApp or Telegram).

Step 3: Launch the Control UI (Your Dashboard)

Now for the fun part. Open your web browser and type in:

http://127.0.0.1:18789/

This address is localhost, meaning it only works on the computer where OpenClaw is running. You'll be greeted by a clean, responsive interface—your control center. Here, you can:

Chat directly with your AI agent.
Monitor connected channels.
Manage your sessions and cron jobs.
View system logs and debug information.

Enter the password you set during onboarding, and you're in.

Why use the Control UI first? While you can set up WhatsApp or Telegram right away, the Control UI is the fastest way to get a live chat. You don't need to link any external accounts—it's ready the moment the gateway is running.

What’s Next?

You now have a working OpenClaw instance. From here, you can:

Connect WhatsApp: Run openclaw channels login --channel whatsapp and scan the QR code with your phone. Your AI can now send and receive WhatsApp messages.
Add Telegram: Create a bot with BotFather, get the token, and link it through the CLI or the web UI.
Explore the CLI: Commands like openclaw gateway status, openclaw message send, and openclaw sessions list let you manage everything from the terminal.
Scale with Nodes: Pair your iPhone or an old Raspberry Pi to extend OpenClaw's reach.

The true power of OpenClaw reveals itself when you begin to automate. Imagine an agent that monitors your Skool community, or one that processes support tickets from email and Slack—no API keys exposed, no data sent to the cloud.

Your AI agent is no longer a chat window. It's an operational partner, and OpenClaw is the key that unlocks its full potential on your personal infrastructure.