OpenClaw on Railway

Create a Railway Project

Open your Railway dashboard and create a new project.

You can deploy OpenClaw using:

For most users, the Railway template is the easiest path. It reduces manual setup and gives you a faster way to get OpenClaw running.

Railway has OpenClaw deployment templates available, including templates that provide setup flow, persistent storage, and OpenClaw gateway management.

Deploy OpenClaw

Deploy OpenClaw from the Railway template, repo, image, or Dockerfile.

For beginners, use the Railway template. For technical users, a custom repo, Docker image, or Dockerfile can work if you want more control.

Once the deployment starts, Railway will build and create the service. After that, open the service settings so you can add variables, storage, and networking.

The easiest option is to use the OpenClaw Railway template. If you are more technical, you can also deploy from your own repo, Docker image, or Dockerfile.

Add Setup Password

Add this Railway variable:

SETUP_PASSWORD protects the setup page.

Use a strong password because /setup is part of the onboarding flow. Do not use something weak like password123, unless your plan is to personally invite chaos into your deployment.

The OpenClaw Railway template documentation lists SETUP_PASSWORD as a required variable for the setup page.

Set Gateway Port

Add this variable:

Railway needs OpenClaw to use the correct port.

OpenClaw’s Railway setup mentions setting required variables, including OPENCLAW_GATEWAY_PORT, and enabling HTTP proxy on port 8080.

This step is important because if the port does not match Railway’s networking setup, your OpenClaw service may deploy but not load correctly.

Add Gateway Token

Add this variable:

OPENCLAW_GATEWAY_TOKEN is the shared secret used to connect with the OpenClaw gateway.

Keep this token private. If it gets exposed, rotate it immediately. This token protects gateway access, so treat it like a password, not like random text you paste into screenshots for strangers to admire.

Mount a Persistent Volume at /data

Add a Railway volume and mount it at:

OpenClaw needs persistent storage for state, sessions, workspace files, and configuration. Without a persistent volume, important setup data may disappear after redeploys or restarts.

Railway volumes provide persistent data for services, and the volume becomes available inside the service at the mount path you choose.

Railway also notes that volumes are used when your application needs to persist data beyond normal deploys and runtime changes.

This is one of the most important steps. Miss it, and your deployment can behave like it has memory loss. Very advanced, very useless.

Set OpenClaw State Directory

Add this variable:

This stores OpenClaw state inside the persistent Railway volume.

It helps preserve setup data after redeploys. Without this, your deployment may lose important state, sessions, or configuration.

The Railway template documentation recommends using:

for persistent OpenClaw state.

Set OpenClaw Workspace Directory

Add this variable:

This stores OpenClaw workspace files inside the /data volume.

Use it together with OPENCLAW_STATE_DIR so your OpenClaw setup can keep workspace data across restarts and redeploys.

The Railway template documentation also lists:

as the workspace path.

Enable HTTP Proxy on Port 8080

Set Railway HTTP proxy to:

Railway public networking exposes your service through HTTP/HTTPS and provides automatic SSL certificates and Railway domains.

OpenClaw Railway setup also mentions enabling HTTP proxy on port 8080, so the proxy port should match your OpenClaw gateway port.

If /openclaw does not load later, this is one of the first things to check.

Open /setup on Your Railway Domain

After deployment, open:

Enter your SETUP_PASSWORD.

Use the setup flow to configure OpenClaw before opening the Control UI. The setup page helps prepare your gateway, workspace, model provider, and basic assistant settings.

The OpenClaw Railway template documentation says to open /setup and complete onboarding before opening /openclaw.

Complete OpenClaw Onboarding

Finish the browser onboarding flow.

During onboarding:

Onboarding prepares your OpenClaw gateway, workspace, auth, and basic assistant setup so the Control UI can work properly.

Do this step carefully. Rushing onboarding is how you create a deployment that technically exists but refuses to be useful, which is a special kind of modern comedy.

Open the OpenClaw Control UI

After onboarding, open:

/openclaw opens the OpenClaw Control UI.

From here, you can manage your assistant, channels, tools, and workflows. Connect using your configured gateway token if required.

OpenClaw Railway documentation says the web Control UI is available after deploying OpenClaw on Railway.

Connect Chat Channels

After the Control UI is ready, connect your channels.

You can connect channels like:

Start with one channel first. Test a simple message before building advanced workflows.

Example flow:

Adding every channel at once is how people create a beautiful multi-platform disaster. Start simple, confirm it works, then expand.