Open the server Apps tab
Select the tutorial-vps VPS, open the Apps tab, and start a new app deployment. Keep sensitive server details hidden before capturing or sharing screenshots.

Add your server credentials to Server Compass
Choose from our template library
Fill in settings and click Deploy
Use the OpenClaw template in Server Compass to deploy a localhost-only AI agent gateway on your VPS, then verify the gateway UI through an SSH tunnel.
Select the tutorial-vps VPS, open the Apps tab, and start a new app deployment. Keep sensitive server details hidden before capturing or sharing screenshots.

Click New App and choose the template deployment path so Server Compass can load the built-in catalog.

Use the template picker search to find OpenClaw in the Server Compass template catalog.

Choose the OpenClaw template. Server Compass fills the gateway service, persistent `.openclaw` data directory, localhost-bound gateway port, bridge port, and generated gateway token.

Confirm the app name and compose services. In this run, the app was named openclaw-demo and used host port 18789.

Review the generated compose settings, confirm the localhost-bound gateway and bridge ports are available, keep the generated gateway token hidden, and click Deploy Now.

Keep the deployment modal open while Server Compass uploads the compose file, builds the OpenClaw gateway image, starts the container, and verifies the stack.

After deployment finishes, return to the Apps tab and confirm the OpenClaw app is marked Running. The gateway is intentionally bound to localhost on the VPS.

Open an SSH tunnel to the gateway port, then load the local tunnel URL in a browser. The OpenClaw gateway UI confirms the service is reachable without exposing it directly to the public internet.

The OpenClaw gateway UI loaded successfully through a local SSH tunnel.
It builds and deploys the OpenClaw gateway container with persistent `.openclaw` storage and localhost-bound gateway and bridge ports.
The tutorial used remote gateway port 18789, bound to 127.0.0.1 on the VPS and accessed through a local SSH tunnel.
A fresh OpenClaw deployment is considered reachable when the gateway UI loads through the SSH tunnel. Token authentication and model-provider configuration depend on your own production credentials.
No. The deployment guide should live on the OpenClaw template detail page and be linked from the reusable template deployment docs page.
Want full control? Here's how to deploy OpenClaw yourself using Docker Compose.
Connect to your remote server using SSH. Make sure you have your server credentials ready.
# Log into your server
ssh root@<your-server-ip>
# If using key-based auth
ssh -i ~/.ssh/my-key root@<your-server-ip>First time? Docker not installed? Run: curl -fsSL https://get.docker.com | sh
Set up a directory to store your Docker configuration.
# Create and navigate to project directory
mkdir -p ~/apps/openclaw
cd ~/apps/openclawAdd this configuration to your docker-compose.yml file:
services:
openclaw-gateway:
build:
context: .
dockerfile: Dockerfile
ports:
- "127.0.0.1:18789:18789"
- "127.0.0.1:18790:18790"
environment:
HOME: "/home/node"
TERM: "xterm-256color"
NODE_ENV: "production"
OPENCLAW_GATEWAY_TOKEN: "<your-gateway-token>"
volumes:
- ./.openclaw:/home/node/.openclaw
- ./.openclaw/workspace:/home/node/.openclaw/workspace
init: true
restart: unless-stopped
command:
- "openclaw"
- "gateway"
- "--allow-unconfigured"
- "--bind"
- "lan"
- "--port"
- "18789"
PORTGateway host port(default: 18789)BRIDGE_PORTBridge host port(default: 18790)GATEWAY_TOKENGateway authentication tokenExecute the deployment and check the container status.
# Deploy the application
docker compose up -d
# Check container health
docker compose ps
# Monitor logs
docker compose logs -f --tail=100Enable network access by updating your firewall rules.
# Allow the application port through firewall
sudo ufw allow 18789/tcp
sudo ufw reload
# Access your app at:
# http://your-server-ip:18789Skip the terminal and deploy OpenClaw with a visual interface. Configure everything with clicks, not commands.
After deploying OpenClaw with Server Compass, complete these steps to finish setup
Open an SSH tunnel via Server Compass to access the Gateway UI
Authenticate with your GATEWAY_TOKEN
Configure Claude API keys in the Control UI
Need help? Check out our documentation for detailed guides.
Common questions about self-hosting OpenClaw
Simply download Server Compass, connect to your VPS, and select OpenClaw from the templates list. Fill in the required configuration and click Deploy. The entire process takes under 3 minutes.
OpenClaw requires a minimum of 256MB RAM. We recommend a VPS with at least 1024MB RAM for optimal performance. Any modern Linux server with Docker support will work.
Yes! Server Compass provides volume mapping that allows you to import existing data. You can also use standard OpenClaw backup and restore procedures.
Server Compass makes updates easy. Simply click the Update button in your deployment dashboard, and the latest OpenClaw image will be pulled and deployed with zero downtime.
OpenClaw is open-source software. You only pay for your VPS hosting (typically $5-20/month) and optionally Server Compass ($29 one-time). No subscription fees or per-seat pricing.

Open-source backend in a single file with realtime database, auth, and file storage

Open-source backend-as-a-service - self-hosted Firebase alternative

Open-source backend framework with dashboard

Full Supabase self-hosted with Kong, GoTrue Auth, Realtime, and Studio
Download Server Compass and deploy OpenClaw to your VPS in under 3 minutes. No Docker expertise required.
Download Server Compass