Getting OpenClaw up and running doesn’t have to be complicated. With the right setup, you can go from cloning the GitHub repository to running your own AI automation platform in less than an hour.

Whether you’re testing OpenClaw on your local machine or deploying it on a VPS for 24/7 availability, this OpenClaw GitHub installation guide walks you through every step. You will learn how to clone the repository, install dependencies, configure the platform, launch the gateway, troubleshoot common issues, and prepare your installation for production.

By the end of this guide, you will have a fully functional OpenClaw installation with a working gateway, an accessible dashboard, and the knowledge to keep it running reliably on your own infrastructure.

If you don’t already have a server, you can deploy OpenClaw on a reliable VPS with Truehost. A VPS keeps your AI agents online around the clock, even when your computer is turned off.

Prerequisites

System Requirements

OpenClaw runs on modest hardware, but a few baseline requirements keep the gateway stable.

• Hardware: 1 vCPU and 1GB RAM minimum for a single-user gateway; 2GB+ RAM recommended for Docker builds
• Operating systems: macOS, Linux, and Windows are all supported
• Node.js: version 24.15+ recommended, with 22.22.3+ or 25.9+ also supported
• Package manager: npm, pnpm, or bun all work for a global CLI install; pnpm is required for source checkouts
• Terminal: any standard shell (bash, zsh, or PowerShell on Windows)

Verify Your Environment

Check your Node.js version before installing anything

node -v

If Node.js is missing or outdated, install it from nodejs.org or a version manager like nvm.

Install pnpm if you plan to build from source.

npm install -g pnpm

Confirm Git is available, since you’ll need it to clone the OpenClaw GitHub repository.

git --version

Clone the OpenClaw GitHub Repository

Clone the repository

Cloning from OpenClaw GitHub gives you the full source, useful if you want to customize the codebase or contribute back.

git clone https://github.com/openclaw/openclaw.git

Navigate to the project directory

cd openclaw

Understand the project structure

A few folders matter most for setup and configuration.

Folder/File	Purpose
src/	Core gateway and CLI source code
extensions/	Bundled channel and skill plugins
config/	Default configuration templates
docs/	Full documentation source
.env.example	Template for environment-based secrets
Dockerfile / docker-compose.yml	Docker install files

Install OpenClaw

Most users don’t need a source checkout at all. If you just want the CLI, skip straight to the global install.

Install dependencies

For a source checkout, this repository is a pnpm workspace, so use pnpm rather than npm.

pnpm install

Build the application

bash

pnpm build
pnpm ui:build

Link OpenClaw globally

If you’d rather skip building from source, install the published package directly.

bash

npm install -g openclaw@latest

Run the onboarding wizard

This is the core of any OpenClaw installation, whether from source or npm.

openclaw onboard --install-daemon

The wizard walks you through gateway setup, workspace creation, channel connections, and skill selection. The --install-daemon flag also installs a background service (systemd on Linux, launchd on macOS) so OpenClaw keeps running after you close the terminal.

---

Configure OpenClaw

Create the initial configuration

Onboarding writes a configuration file automatically. You can also edit it directly.

~/.openclaw/openclaw.json

A minimal config looks like this:

json

{
  "agents": {
    "defaults": {
      "model": "<provider>/<model-id>"
    }
  }
}

Configure AI provider API keys

OpenClaw supports Anthropic, OpenAI, Google Gemini, xAI Grok, OpenRouter, GitHub Copilot, MiniMax, and any OpenAI- or Anthropic-compatible endpoint. Add keys through the onboarding wizard, or set them as environment variables.

Configure the gateway

The gateway defaults to loopback binding for security, meaning only your local machine can reach it. Change this only if you understand the security trade-offs of exposing it more broadly.

Configure the workspace

Your OpenClaw configuration includes a workspace root where the agent stores files and context.

~/.openclaw/workspace

This path is configurable via agents.defaults.workspace if you’d rather store it elsewhere.

Configure optional environment variables

Use environment variables instead of hardcoding values whenever a setting is sensitive (API keys, gateway tokens) or changes between environments (development vs. production). Store them in a .env file at the project root, based on .env.example.

Launch OpenClaw

Development mode

Use this while actively editing source code, since it auto-reloads on changes.

bash

pnpm gateway:watch

Production mode

For a standard install, start the gateway service directly.

bash

openclaw gateway start

Start background services

The onboarding wizard’s --install-daemon flag already sets this up, keeping the gateway running as a persistent background service. Use this for any long-running deployment, including a VPS.

---

Verify Your Installation

Run diagnostics

bash

openclaw doctor

This checks your configuration, dependencies, and channel setup for common issues.

Check gateway status

bash

openclaw gateway status

Expected output confirms the gateway is running and shows its port, typically 18789.

Verify daemon status

If you installed the background service, openclaw gateway status also reflects its state. On Linux, you can cross-check with systemd directly.

bash

systemctl --user status openclaw-gateway.service

Test the dashboard

bash

openclaw dashboard

This opens the Control UI in your browser.

Run your first AI request

bash

openclaw agent --message "Hello from OpenClaw" --thinking high

A successful response confirms your provider API key and model configuration are working correctly.

Access the OpenClaw Dashboard

Default dashboard URL

By default, the OpenClaw dashboard is available at:

http://127.0.0.1:18789

Dashboard overview

• Agents: manage your configured agents and their assigned models
• Channels: view and connect messaging platforms like Telegram, Slack, or Discord
• Logs: review gateway activity and troubleshoot errors
• Configuration: edit settings without touching the JSON file directly

Change the default dashboard port

Update the gateway’s bind and port settings in openclaw.json, then restart the gateway for the change to take effect.

Deploy OpenClaw on a VPS

Why use a VPS

A VPS keeps OpenClaw running continuously, independent of your personal laptop or desktop uptime. This matters if you’re connecting messaging channels that need to respond around the clock. If you’d rather skip the server setup entirely, Truehost OpenClaw Hosting provisions a ready-to-run gateway on infrastructure built for this exact workload.

Install on Ubuntu

The same install commands from earlier apply directly on an Ubuntu server.

bash

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

Install on macOS

If you’re testing locally before deploying to a VPS, macOS uses the same install script as Linux.

bash

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

The --install-daemon flag sets up a launchd service, so the gateway keeps running in the background even after you close the terminal.

Install on Windows

Windows uses a dedicated PowerShell install script instead of the shell script.

powershell

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

Run onboarding the same way once the install finishes.

powershell

openclaw onboard --install-daemon

Windows also has a native Hub companion app if you’d rather manage the gateway through a desktop interface instead of the termin

Configure the firewall

Only open the ports you actually need. If you’re accessing the dashboard through an SSH tunnel or Tailscale, you don’t need to open the gateway port publicly at all.

bash

sudo ufw allow OpenSSH
sudo ufw allow 443/tcp
sudo ufw enable

Configure Nginx or Caddy as a reverse proxy

If you want a public URL for the dashboard, put a reverse proxy in front of the gateway rather than exposing port 18789 directly.

nginx

server {
    listen 80;
    server_name yourdomain.com;

    location / {
        proxy_pass http://127.0.0.1:18789;
        proxy_set_header Host $host;
        proxy_set_header X-Real-IP $remote_addr;
    }
}

Caddy handles this with less configuration:

yourdomain.com {
    reverse_proxy 127.0.0.1:18789
}

Enable HTTPS using Let’s Encrypt

With Nginx, Certbot automates certificate issuance and renewal.

bash

sudo certbot --nginx -d yourdomain.com

Caddy issues and renews HTTPS certificates automatically, with no extra command needed.

Run OpenClaw as a systemd service

The onboarding wizard installs this automatically when you use --install-daemon. To review or edit the unit file:

bash

systemctl --user edit openclaw-gateway.service

Automatically restart OpenClaw after a reboot

Enable the service so it starts on boot without manual intervention.

bash

systemctl --user enable openclaw-gateway.service

For extra resilience on small VMs, set explicit restart behavior in the unit override:

ini

[Service]
Restart=always
RestartSec=2
TimeoutStartSec=90

Running this whole setup on dependable infrastructure matters more than any single config file. Truehost OpenClaw Hosting is built specifically for this kind of always-on gateway, so you skip the manual server tuning and get straight to configuring your agent.

Common Installation Errors

Error	Fix
Incorrect Node.js version	Install Node 24.15+, 22.22.3+, or 25.9+; check with node -v
pnpm installation failures	Reinstall with npm install -g pnpm and retry
Dependency conflicts	Delete node_modules and lockfile artifacts, then reinstall
Missing API keys	Re-run openclaw onboard to add provider credentials
Gateway startup failures	Run openclaw doctor to surface the specific misconfiguration
Port already in use	Change the gateway port in openclaw.json, or stop the conflicting process
Permission errors	Check ownership of ~/.openclaw; avoid mixing root and user installs
Build failures	Confirm you’re using pnpm (not npm) for source builds, and that Node meets the minimum version

Security Best Practices

• Protect the dashboard. Keep the gateway on loopback and access it through an SSH tunnel or Tailscale rather than exposing it publicly.
• Secure API keys. Store them as environment variables or in the encrypted auth-profile store, never hardcoded in source files.
• Restrict agent permissions. Use sandboxing for non-main sessions, especially in group or multi-user channels.
• Configure firewall rules. Only open ports you actively need; keep the gateway port closed to the public internet where possible.
• Store secrets using environment variables. This keeps sensitive values out of version control and config files you might share.
• Keep dependencies updated. Outdated packages are one of the most common sources of vulnerabilities in any self-hosted tool.

Updating OpenClaw

Update a GitHub installation

bash

git pull
pnpm install
pnpm build

Update a CLI installation

bash

openclaw update

Switch release channels if needed:

bash

openclaw update --channel stable

Verify the update completed successfully

bash

openclaw doctor

Run this OpenClaw update guide step every time, since it catches migration issues before they become runtime errors

Managing Your Installation

• View logs: check the Logs tab in the dashboard, or tail the gateway’s log output directly
• Restart services: openclaw gateway stop followed by openclaw gateway start
• Stop OpenClaw: openclaw gateway stop
• Reset configuration: re-run openclaw onboard to overwrite existing settings
• Back up configuration files: copy ~/.openclaw/openclaw.json and the workspace directory to a safe location regularly
• Uninstall OpenClaw: follow the official uninstall steps to remove the daemon, config, and workspace cleanly

Install OpenClaw with Docker (Optional)

Install with Docker

From the repository root:

bash

./scripts/docker/setup.sh

This builds the gateway image locally and runs onboarding automatically, including generating a gateway token.

Docker Compose example

bash

docker compose up -d openclaw-gateway

Once running, open the Control UI at http://127.0.0.1:18789 and paste the token from your .env file.

When Docker is a better choice than a native installation

Docker makes sense for an isolated, disposable environment, or when you want consistent behavior across multiple hosts without managing Node.js versions manually. For a straightforward single-VPS deployment, a native OpenClaw Ubuntu installation is often simpler to manage day to day

Production Best Practices

• Monitor CPU and memory usage, especially during Docker image builds
• Keep Node.js updated to a supported version
• Schedule regular backups of your config and workspace directories
• Rotate API keys periodically, particularly after any suspected exposure
• Update dependencies regularly to pick up security patches
• Restrict public dashboard access to trusted networks only
• Monitor logs for repeated errors or unexpected channel activity
• Test backups periodically by restoring them on a separate machine

FAQs

Ready to Build with OpenClaw?

You have gone from cloning the OpenClaw GitHub repository to configuring the gateway, verifying your installation, and deploying it on a VPS with a reverse proxy and HTTPS. That’s the full lifecycle, covered end to end.

From here, start connecting channels, experiment with different AI providers, and build out the workflows your agent will handle daily. For anything meant to run continuously, deploying on a VPS beats relying on a personal machine that sleeps, reboots, or loses its network connection.

Ready to deploy? Truehost OpenClaw Hosting gives your OpenClaw VPS setup the uptime, resources, and reliability a production gateway needs, so your assistant stays online whenever your team needs it. Get OpenClaw hosting from Truehost running today and deploy with confidence.