How to deploy OpenClaw on a VPS for 24/7 uptime

For a real always-on VPS, the current official production recommendation is openclaw-ansible, not a hand-built shell script. It is the shortest official path to a hardened deployment with firewalling, Docker isolation, Tailscale access, and systemd auto-start.

Who this is for

• You want OpenClaw running continuously on a remote Linux host.
• You want a production-style setup, not a local dev install.
• You want a safer deployment than binding the dashboard directly to the public internet.

What you need before you start

• A Linux VPS you control.
• SSH access and sudo on that host.
• A plan for secure remote access, such as Tailscale or SSH tunneling.

Step-by-step setup

Step 1: SSH into the VPS

Expected result: you have a shell on the Linux VPS where OpenClaw will run.

Step 2: Run the official Ansible installer

The official Ansible page describes this as the recommended way to deploy OpenClaw to production servers.

Step 3: Finish the guided install

During the installer flow, keep the security-first defaults unless you have a specific reason to change them. The Ansible docs emphasize firewall-first access and Tailscale exposure instead of public dashboard exposure.

Step 4: Verify the gateway service on the VPS

Expected result: the gateway service is running after the installer completes.

Step 5: Open the dashboard through a safe access path

Generate the dashboard link on the VPS:

Then either use Tailscale Serve or an SSH tunnel. For a plain SSH tunnel:

Open http://127.0.0.1:18789/ locally after the tunnel is up.

Verify it worked

1. openclaw gateway status reports a healthy runtime on the VPS.
2. The dashboard loads through Tailscale or your SSH tunnel.
3. The gateway survives a normal disconnect from your SSH session because it is running as a managed service.

Common problems and fixes

The service does not stay up after deploy

Use the official troubleshooting commands:

The dashboard is reachable but not secure

Do not expose the Control UI directly to the public internet. The dashboard docs explicitly call it an admin surface and recommend localhost, Tailscale Serve, or an SSH tunnel.

The gateway port is already in use

Check the runtime and port conflict state with:

FAQ

Why use the Ansible installer instead of the regular Linux install on a VPS?

Because the official deployment docs recommend openclaw-ansible for production servers and describe it as the source of truth for secure deployment.

Can I still manage OpenClaw with the normal CLI after deployment?

Yes. Once deployed, the same openclaw CLI commands such as gateway status, dashboard, doctor, and config helpers still apply on the host.

Should the dashboard ever be public?

No. The dashboard docs explicitly say it is an admin surface and should stay behind localhost, Tailscale, or an SSH tunnel.

Official sources

• Ansible Installation
• Dashboard
• Gateway CLI
• Gateway Troubleshooting