Railway MCP Server

Skill by ara.so — MCP Skills collection.

Railway MCP Server is an official Model Context Protocol server that enables AI assistants to interact with Railway infrastructure. It's now bundled directly into the Railway CLI, allowing you to manage Railway projects, deployments, databases, environment variables, and services programmatically.

Installation

Install Railway CLI

The Railway MCP server is bundled with the Railway CLI (v4.0.0+):

bash <(curl -fsSL https://railway.com/install.sh)

Configure MCP Client

Automatically configure supported MCP clients (Claude Desktop, Cline, etc.):

railway mcp install

For remote (hosted) MCP server instead of local stdio:

railway mcp install --remote

Manual Configuration

Add to your MCP client configuration (e.g., claude_desktop_config.json):

{
  "mcpServers": {
    "railway": {
      "command": "railway",
      "args": ["mcp"]
    }
  }
}

Authentication

railway login

This opens a browser for authentication and stores credentials locally.

Core Capabilities

Project Management

List all projects:

railway projects

Create a new project:

railway init

Link current directory to a project:

railway link

Switch projects:

railway project

Service Deployment

Deploy current directory:

railway up

Deploy with specific Dockerfile:

railway up --dockerfile ./Dockerfile.prod

View deployment status:

railway status

List all services:

railway service

Environment Variables

Set environment variable:

railway variables set KEY=value

Set multiple variables:

railway variables set API_KEY=$API_KEY DB_HOST=postgres.railway.internal

List all variables:

railway variables

Delete a variable:

railway variables delete KEY

Database Management

Add a database (Postgres, MySQL, Redis, MongoDB):

railway add

Connect to database shell:

railway connect postgres

Get database connection string:

railway variables | grep DATABASE_URL

Logs and Monitoring

View service logs:

railway logs

Follow logs in real-time:

railway logs --follow

View logs for specific deployment:

railway logs --deployment <deployment-id>

Domains

Add custom domain:

railway domain

Generate Railway subdomain:

railway domain --generate

MCP Server Usage

When running through MCP, the Railway server exposes tools that AI assistants can invoke:

Available MCP Tools

railway_projects_list - List all Railway projects
railway_project_create - Create a new project
railway_services_list - List services in a project
railway_service_create - Create a new service
railway_deploy - Deploy a service
railway_deployments_list - List deployments
railway_deployment_status - Get deployment status
railway_variables_list - List environment variables
railway_variables_set - Set environment variables
railway_variables_delete - Delete environment variables
railway_logs_get - Retrieve service logs
railway_databases_add - Add a database service
railway_domains_list - List domains
railway_domains_add - Add a custom domain

Example MCP Interactions

AI Assistant Usage:

When an AI assistant has Railway MCP configured, users can make natural language requests:

User: "Deploy my Node.js app to Railway"

The AI will:

Check if project is linked (railway_project_status)
Create project if needed (railway_project_create)
Deploy the service (railway_deploy)
Monitor deployment status (railway_deployment_status)
Return the deployment URL

User: "Add Postgres database and set DATABASE_URL"

The AI will:

Add Postgres plugin (railway_databases_add)
Retrieve connection string (railway_variables_list)
Set DATABASE_URL if needed (railway_variables_set)

Configuration Files

railway.json (Project Config)

{
  "$schema": "https://railway.app/railway.schema.json",
  "build": {
    "builder": "NIXPACKS",
    "buildCommand": "npm run build"
  },
  "deploy": {
    "startCommand": "npm start",
    "healthcheckPath": "/health",
    "restartPolicyType": "ON_FAILURE",
    "restartPolicyMaxRetries": 10
  }
}

railway.toml (Multi-Service Config)

[build]
builder = "NIXPACKS"
buildCommand = "npm run build"

[deploy]
startCommand = "npm start"
healthcheckPath = "/health"
healthcheckTimeout = 300
restartPolicyType = "ON_FAILURE"

[[services]]
name = "frontend"
source = "./frontend"

[[services]]
name = "backend"
source = "./backend"

Common Patterns

Full Stack Deployment

# Initialize project
railway init

# Add Postgres database
railway add --database postgres

# Deploy backend service
cd backend
railway up

# Deploy frontend service
cd ../frontend
railway up

# Set environment variables
railway variables set \
  BACKEND_URL=https://backend.railway.app \
  DATABASE_URL=$DATABASE_PRIVATE_URL

CI/CD Integration

GitHub Actions example:

name: Deploy to Railway
on:
  push:
    branches: [main]

jobs:
  deploy:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v3
      
      - name: Install Railway
        run: bash <(curl -fsSL https://railway.com/install.sh)
      
      - name: Deploy
        run: railway up --service backend
        env:
          RAILWAY_TOKEN: ${{ secrets.RAILWAY_TOKEN }}

Environment-Specific Variables

# Production environment
railway environment production
railway variables set NODE_ENV=production API_URL=https://api.prod.example.com

# Staging environment
railway environment staging
railway variables set NODE_ENV=staging API_URL=https://api.staging.example.com

Troubleshooting

MCP Server Not Found

If railway mcp command fails:

# Check Railway CLI version (must be 4.0.0+)
railway --version

# Upgrade CLI
bash <(curl -fsSL https://railway.com/install.sh)

Authentication Issues

# Re-authenticate
railway logout
railway login

# Verify authentication
railway whoami

Deployment Failures

# View detailed logs
railway logs --deployment <deployment-id>

# Check build logs
railway logs --build

# Verify service configuration
railway status

MCP Client Configuration Issues

Remove legacy @railway/mcp-server npm package configurations:

// OLD - Remove this
{
  "mcpServers": {
    "railway": {
      "command": "npx",
      "args": ["-y", "@railway/mcp-server"]
    }
  }
}

// NEW - Use this
{
  "mcpServers": {
    "railway": {
      "command": "railway",
      "args": ["mcp"]
    }
  }
}

Connection String Access

# Private network URL (for internal services)
railway variables | grep PRIVATE_URL

# Public URL (for external access)
railway variables | grep DATABASE_URL

Documentation

Official Docs: https://docs.railway.com
CLI MCP Docs: https://docs.railway.com/cli/mcp
Railway CLI Reference: https://docs.railway.com/cli/quick-start
Railway Templates: https://railway.app/templates

railway-mcp-server