How to Install OpenClaw on a VPS?

A VPS (Virtual Private Server) is the most popular way to run OpenClaw 24/7. You get a dedicated server with a static IP address, guaranteed uptime, and full root access.

Any VPS provider works: DigitalOcean, Hetzner, Hostinger, Vultr, Linode, AWS Lightsail. Choose a plan with at least 2 GB RAM, 1 vCPU, and 20 GB SSD. Ubuntu 22.04 LTS is recommended. Prices start at $5-10/month.

Connect via SSH and run the install script. It sets up Node.js, downloads OpenClaw, and launches the onboarding wizard. Choose your AI model, enter the API key, and connect Telegram (or another messenger).

After setup, configure a systemd service for auto-restart. Set up a basic firewall with ufw — allow SSH (22), your Gateway port (18789), and block everything else. For the Dashboard, use Nginx as a reverse proxy with Let's Encrypt SSL.

Create a non-root user to run OpenClaw for security. Never run production services as root. The install script can create the user and service automatically if you choose.

For cost optimization, a $5-7/month VPS handles a single agent with cloud models (Claude, GPT) easily. If you want local models via Ollama, you'll need a more powerful server — 16 GB+ RAM and ideally a GPU.

# Connect to your VPS
ssh root@your-server-ip

# Install OpenClaw
curl -fsSL https://openclaw.sh/install | bash
openclaw onboard

# Set up firewall
sudo ufw allow ssh
sudo ufw allow 18789
sudo ufw enable