Railway MCP server

What the model can do

Railway hosting: projects with their services and environments, a service's recent deployments with status and URL, and one deployment's status. When asked, deploys or redeploys a service, restarts, rolls back, stops or cancels a deployment, creates, renames, reconfigures or deletes a service, and creates or renames a project. Environment variables and logs are never read or written.

Reads

• railway_projectsRailway projects the token can see (the workspace's when a workspace ID is set): id, name, description, created and updated dates.
• railway_projectOne Railway project with its services and environments (names and ids), which railway_deployments needs.
• railway_deploymentsRecent deployments of one Railway service in one environment, newest first: status (BUILDING, DEPLOYING, SUCCESS, FAILED, CRASHED, REMOVED…
• railway_deploymentOne Railway deployment's status, dates, URL and the service, environment and project it belongs to.

Acts, when you ask

These change something in Railway. The model is told to use them only when you ask for exactly that, and each one is granted by name: per team on a shared connection, by the owner on a personal one.

• railway_deploy_serviceDeploy one Railway service in one environment now.
• railway_redeploy_serviceRedeploy a Railway service's latest deployment in one environment with its existing commit (no new commits are pulled…
• railway_restart_deploymentRestart a running Railway deployment without rebuilding it.
• railway_rollback_deploymentRoll a Railway service back to an earlier deployment, which becomes the running one.
• railway_stop_deploymentStop a running Railway deployment.
• railway_cancel_deploymentCancel a Railway deployment that is still building or queued.
• railway_redeploy_deploymentRedeploy one specific Railway deployment (rebuilds it and makes it the running one).
• railway_create_serviceCreate a service in a Railway project: empty, from a GitHub repo (owner/repo, optional branch) or from a Docker image.
• railway_update_service_settingsChange a Railway service's build and deploy settings in one environment: start or build command, root directory, Dockerfile path, watch patterns…
• railway_update_serviceRename a Railway service, or change its icon.
• railway_delete_servicePermanently delete a Railway service and all its deployments, in every environment.
• railway_create_projectCreate an empty, private Railway project, in the connection's workspace when one is set (or the workspace given).
• railway_update_projectRename a Railway project or change its description.

Connect it

First connect Railway on Dashboard → Connections, then add Memtro to your client once. The Railway tools appear alongside the tools of everything else you have connected; nothing is configured per connector.

claude plugin marketplace add memtro-org/claude-memtro
claude plugin install memtro@memtro --config api_key=sk-memtro-...

{
  "mcpServers": {
    "memtro": {
      "url": "https://platform.memtro.com/api/mcp",
      "headers": {
        "Authorization": "Bearer sk-memtro-..."
      }
    }
  }
}

{
  "mcp": {
    "memtro": {
      "type": "remote",
      "url": "https://platform.memtro.com/api/mcp",
      "headers": {
        "Authorization": "Bearer sk-memtro-..."
      }
    }
  }
}

Who may use it

The key identifies the person, and the server offers that person the tools they may use everywhere else in Memtro. A shared Railway connection is used by the teams an administrator granted it to, tool by tool, on the Permissions page; a personal connection is used by its owner alone, and its tools that act stay off until the owner allows them. See Teams and permissions and Connectors.

Ask it like this

• › Did the last deploy of the API service succeed?
• › Which services are in the billing project?
• › Has the worker service crashed in production this week?
• › What URL is the staging web service on?
• › Roll the API service in production back to the previous deployment

Questions

Is this an official Railway MCP server?

No. It is Memtro's MCP server with Memtro's Railway connector behind it: the connector reads Railway through its API with the credentials you connect on Dashboard → Connections, and Railway publishes nothing here. The same endpoint carries every other system you connect and Memtro's memory tools.

What does it cost?

Nothing on top of Memtro. One price per organisation by its number of users, from $50 a month for up to 10 users after a 7-day free trial, with every connector, the MCP server, jobs and memory included. You bring your own model keys.

Can it change anything in Railway?

Only when you ask, and only what it has been granted: deploy a service in an environment, optionally at a given commit (starts a build and billable compute); redeploy a service's latest deployment with its existing commit; restart a running deployment without rebuilding (brief downtime); roll a service back to an earlier deployment; stop a running deployment (the service goes offline); cancel a deployment that is building or queued; redeploy one specific earlier deployment; create a service in a project, empty or from a GitHub repo or Docker image; change a service's build and deploy settings in one environment (commands, replicas, region, cron, healthcheck); rename a service or change its icon; delete a service and all its deployments (permanent); create an empty project; rename or re-describe a project. Every other tool reads, searches or lists. An administrator grants each acting tool by name per team, and a personal connection acts only after its owner allows it.

Which clients work?

Any MCP client that speaks Streamable HTTP: Claude Code, Cursor, Claude Desktop, OpenCode, Codex, Zed, Windsurf, Cline and the Anthropic and OpenAI SDKs. There is no local process to run; the server is remote and authenticated with your key.

Other MCP servers · Engineering and operations

• GitHub MCP server
• AWS CloudWatch MCP server
• Sentry MCP server
• ServiceNow MCP server
• Okta MCP server
• Twilio MCP server