How to Deploy OpenClaw on Northflank - One-Click Stack Guide

Northflank is the lightest-weight hosted deployment path in the current docs. The official page is intentionally short because the value proposition is simple: one-click stack, hosted gateway, persistent volume, and a public OpenClaw URL where you can log into the Control UI without touching a terminal on the server.

What the official docs support

The docs describe a one-click template that runs OpenClaw for you, asks for OPENCLAW_GATEWAY_TOKEN, and exposes the Control UI at /openclaw. Persistent storage is mounted at /data so openclaw.json, auth-profiles, channel state, sessions, and workspace content survive redeploys.

This is a good fit when you want a hosted platform but do not need a lot of custom infrastructure ceremony. It is especially appealing for small teams that want something easier than a raw VPS yet still want persistent state and a normal OpenClaw dashboard instead of a locked-down vendor-specific UI.

What you need first

• A Northflank account
• A strong random value for OPENCLAW_GATEWAY_TOKEN
• A plan for the first messaging channel you want to connect
• Comfort with a public service URL fronting the Control UI

Recommended setup flow

There is not much to this path, which is the point. The operator work is mostly around choosing a good token and knowing where the persistent volume lives.

1. Open the Deploy OpenClaw template on Northflank and create an account if needed. The docs describe this as the easiest no-terminal-on-the-server path for OpenClaw hosting.
2. Click Deploy OpenClaw now, set OPENCLAW_GATEWAY_TOKEN to a strong random value, and deploy the stack. That token is the control-plane secret for the dashboard unless you later move to password auth.
3. Wait for the stack to finish, then open View resources and click into the OpenClaw service. The docs say the public OpenClaw URL serves the app under /openclaw, which is the path you should bookmark for the Control UI.
4. Confirm the persistent volume is in place at /data. That is where your config, sessions, provider state, and workspace survive redeploys, and it is the main reason the template is more than a throwaway demo.
5. Once the dashboard is reachable, add your first channel from the Control UI or use openclaw onboard over SSH if you have shell access enabled. Telegram is called out in the docs as the fastest first connection.

Set OPENCLAW_GATEWAY_TOKEN=<strong-random-value>
Deploy the official Northflank OpenClaw stack
Open the service at /openclaw

Access, safety, and operational notes

This guide is built around the public Control UI path, so the most important access choice is a strong gateway token. Because the docs keep the rest of the deployment simple, that shared secret does a lot of real security work and deserves proper handling.

The volume mount at /data is the key operational detail. Without it, this would just be another stateless container demo. With it, the deployment can survive normal redeploys while keeping the things that actually make your assistant useful.

If you plan to hand this deployment to another teammate later, save two things on day one: the public service URL and the gateway token. The official setup is simple enough that losing either one creates more friction than the actual deploy.

How to verify it is working

Open the Control UI at /openclaw, confirm the token works, then add one channel and send a single message through it. That proves both the platform deployment and the OpenClaw layer above it are functioning the way the short official guide expects.

Common gotchas

• The Control UI lives at /openclaw, not only at the service root
• OPENCLAW_GATEWAY_TOKEN is required and should be set deliberately
• The persistent volume at /data is the difference between a real deployment and a disposable one

If you want the operator version with tighter rollout checklists, safer defaults, and more production patterns, The OpenClaw Playbook is the easiest shortcut.