How to Install OpenClaw in 2026: The Complete Setup Guide

OpenClaw has quickly become the go-to self-hosted AI assistant gateway for developers in 2026. It connects your favorite messaging platforms — WhatsApp, Telegram, Discord, iMessage — directly to powerful AI coding agents, all running on your own hardware. No cloud middlemen. No data leaks.

In this guide, we’ll walk through every step to get OpenClaw installed and running from scratch.

Prerequisites

Before we start, make sure you have the following:

Node.js 22.12.0 or later (we recommend Node.js 24 LTS)
An LLM API key from Anthropic, OpenAI, or Google
A dedicated machine or VPS — do NOT install OpenClaw on your primary work machine for security reasons
Windows users: WSL2 (Windows Subsystem for Linux) is required

Check Your Node.js Version

node -v

If you’re below v22.12.0, upgrade via nvm or download from nodejs.org.

# Using nvm (recommended)
nvm install 24
nvm use 24

Step 1: Install OpenClaw via NPM

The fastest way to install OpenClaw is through NPM’s global package.

macOS / Linux

npm install -g openclaw@latest

Windows (PowerShell — runs inside WSL2)

# First, ensure WSL2 is installed
wsl --install

# Inside your WSL2 terminal:
npm install -g openclaw@latest

Alternative: Shell Script (macOS / Linux)

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

Verify the installation:

openclaw --version

You should see output like openclaw v3.x.x.

Step 2: Run the Onboarding Wizard

OpenClaw ships with a built-in onboarding wizard that configures authentication, gateway settings, and messaging channels all in one flow.

openclaw onboard --install-daemon

The wizard will guide you through:

Authentication Setup — Create a secure passphrase for the gateway
AI Provider Selection — Choose between Anthropic (Claude), OpenAI, or Google
API Key Configuration — Paste your LLM API key
Channel Setup — Connect your first messaging platform (Telegram is the easiest to start with)

Step 3: Connect Your First Channel (Telegram)

Telegram is the simplest channel to configure. Here’s the process:

Open Telegram and search for @BotFather
Send /newbot and follow the prompts to create your bot
Copy the bot token BotFather gives you
Get your Telegram User ID (send a message to @userinfobot)

When the onboarding wizard asks for channel configuration:

? Select a channel to connect: Telegram
? Enter your Telegram Bot Token: 7123456789:AAHxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
? Enter your Telegram User ID: 123456789

Step 4: Verify the Gateway

Once onboarding is complete, check that everything is running:

openclaw gateway status

You should see output similar to:

┌─────────────────────────────────────┐
│  OpenClaw Gateway Status            │
├─────────────────────────────────────┤
│  Status:    ● Running               │
│  Uptime:    2m 34s                  │
│  Channels:  Telegram (connected)    │
│  AI Model:  claude-3.5-sonnet       │
│  Version:   3.2.1                   │
└─────────────────────────────────────┘

Step 5: Send Your First Message

Open Telegram and send a message to your new bot:

Hey OpenClaw, what can you do?

The AI will respond with its capabilities — file management, command execution, web search, and more.

Step 6: Access the Dashboard

OpenClaw also provides a browser-based control panel:

openclaw dashboard

This opens a local web UI (usually at http://localhost:3377) where you can:

Chat with your AI assistant directly
Monitor gateway activity
Configure tools and skills
View session history

Troubleshooting Common Issues

”Command not found: openclaw”

Make sure your global NPM bin directory is in your $PATH:

# Check your npm global prefix
npm config get prefix

# Add to your shell profile (e.g. ~/.zshrc or ~/.bashrc)
export PATH="$(npm config get prefix)/bin:$PATH"
source ~/.zshrc

Port Already in Use

If port 3377 is occupied:

openclaw gateway start --port 3400

WSL2 Networking Issues on Windows

Ensure your WSL2 distribution has proper network access:

# Inside WSL2
ping 8.8.8.8
curl -I https://registry.npmjs.org

Security Best Practices

Since OpenClaw can execute commands and access files, security is paramount:

Use a dedicated machine or VPS — never your primary workstation
Bind to localhost only — use a VPN for remote access
Run as a non-root user — create a dedicated openclaw system user
Use burner accounts for connected messaging platforms

# Create a dedicated user (Linux)
sudo useradd -m -s /bin/bash openclaw
sudo su - openclaw

# Install and run as this user
npm install -g openclaw@latest
openclaw onboard --install-daemon

What’s Next?

Now that OpenClaw is installed, you’re ready to configure custom tools, create skills, and connect additional channels. Check out our other OpenClaw tutorials for advanced setups like browser automation, multi-agent routing, and Docker deployments.

Need help deploying OpenClaw for your team? Contact our engineering team — we specialize in self-hosted AI infrastructure.