Cloudflare Workers Deployment

Deploy FastMCP to Cloudflare Workers for global edge distribution

Deploy your FastMCP server to Cloudflare Workers for ultra-low latency worldwide. The edge runtime allows your MCP server to run close to your users with automatic global distribution.

EdgeFastMCP vs FastMCP

Cloudflare Workers requires the edge-compatible EdgeFastMCP class:

Feature	FastMCP	EdgeFastMCP
Runtime	Node.js	Edge (V8 isolates)
Start method	server.start({ port })	export default server
Transport	stdio, httpStream, SSE	HTTP Streamable only
Sessions	Stateful or stateless	Stateless only
File system	Yes	No
OAuth/Authentication	Built-in authenticate option	Use Hono middleware (built-in planned)
Custom routes	server.getApp()	server.getApp()

EdgeFastMCP is optimized for stateless edge environments with no filesystem access. Use fetch APIs for external data.

Step-by-Step Deployment

1. Install Dependencies

   Install FastMCP and Zod in your Cloudflare Workers project:

   npm install fastmcp zod
   
2. Create Your MCP Server

   Create src/index.ts with the EdgeFastMCP import:

   src/index.ts

   import { EdgeFastMCP } from "fastmcp/edge";
   import { z } from "zod";
   
   const server = new EdgeFastMCP({
     name: "My Edge Server",
     version: "1.0.0",
     description: "MCP server running on Cloudflare Workers",
   });
   
   // Add tools, resources, prompts as usual
   server.addTool({
     name: "greet",
     description: "Greet someone",
     parameters: z.object({
       name: z.string(),
     }),
     execute: async ({ name }) => {
       return `Hello, ${name}! Served from the edge.`;
     },
   });
   
   // Export the server as default (required for Cloudflare Workers)
   export default server;
   

   The export default server statement is required for Cloudflare Workers to recognize your handler.

3. Configure wrangler.toml

   Create a wrangler.toml configuration file:

   wrangler.toml

   name = "my-mcp-server"
   main = "src/index.ts"
   compatibility_date = "2024-01-01"
   

   Basic Configuration

   name = "my-mcp-server"
   main = "src/index.ts"
   compatibility_date = "2024-01-01"
   
4. Deploy to Cloudflare

   Deploy your MCP server with a single command:

   npx wrangler deploy
   

   Your server will be available at:

   • MCP endpoint: https://my-mcp-server.workers.dev/mcp
   • Health check: https://my-mcp-server.workers.dev/health

Complete Example

Here’s a full example based on src/examples/edge-cloudflare-worker.ts:

src/index.ts

import { EdgeFastMCP } from "fastmcp/edge";
import { z } from "zod";

const server = new EdgeFastMCP({
  description: "An MCP server running on Cloudflare Workers",
  name: "CloudflareWorkerMCP",
  version: "1.0.0",
});

// Simple tool
server.addTool({
  name: "greet",
  description: "Greet someone by name",
  parameters: z.object({
    name: z.string().describe("The name to greet"),
  }),
  execute: async ({ name }) => {
    return `Hello, ${name}! This response is from a Cloudflare Worker.`;
  },
});

// Tool with structured content
server.addTool({
  name: "get_weather",
  description: "Get weather information for a location",
  parameters: z.object({
    location: z.string().describe("The city or location"),
  }),
  execute: async ({ location }) => {
    // Call external API using fetch
    return {
      content: [
        {
          type: "text",
          text: `Weather for ${location}:\n- Temperature: 72°F\n- Conditions: Sunny\n- Humidity: 45%`,
        },
      ],
    };
  },
});

// Static resource
server.addResource({
  uri: "info://server",
  name: "Server Info",
  description: "Information about this MCP server",
  mimeType: "text/plain",
  load: async () => {
    return "This is a FastMCP server running on Cloudflare Workers edge runtime.";
  },
});

// Prompt template
server.addPrompt({
  name: "analyze_code",
  description: "Generate a prompt to analyze code",
  arguments: [
    { name: "language", description: "Programming language", required: true },
    { name: "focus", description: "What to focus on (optional)", required: false },
  ],
  load: async (args) => {
    const focus = args.focus ? ` focusing on ${args.focus}` : "";
    return {
      messages: [
        {
          role: "user",
          content: {
            type: "text",
            text: `Please analyze the following ${args.language} code${focus}:`,
          },
        },
      ],
    };
  },
});

export default server;

Adding Custom Routes

Access the underlying Hono app to add custom HTTP endpoints:

const app = server.getApp();

// Add a landing page
app.get("/", (c) => c.html("<h1>Welcome to my MCP server</h1>"));

// Add REST API endpoints
app.get("/api/status", (c) => c.json({ status: "ok", edge: true }));

// Handle POST requests
app.post("/api/data", async (c) => {
  const body = await c.req.json();
  return c.json({ received: body });
});

export default server;

Environment Variables and Secrets

1. Add Environment Variables

   Define variables in wrangler.toml:

   [vars]
   API_ENDPOINT = "https://api.example.com"
   CACHE_TTL = "3600"
   
2. Add Secrets

   Use wrangler CLI for sensitive data:

   npx wrangler secret put API_KEY
   npx wrangler secret put DATABASE_URL
   
3. Access in Code

   TypeScript doesn’t have direct access to env in EdgeFastMCP. Use Hono middleware:

   const app = server.getApp();
   
   app.use("*", async (c, next) => {
     // Access environment variables
     const apiKey = c.env.API_KEY;
     await next();
   });
   

Never commit secrets to wrangler.toml. Always use wrangler secret put for sensitive values.

Edge Runtime Limitations

When running on Cloudflare Workers, be aware of these constraints:

• No filesystem access: Use KV, R2, or fetch APIs instead
• Stateless by default: Each request is handled independently
• V8 isolates: Limited to JavaScript/TypeScript runtime
• CPU time limits: Maximum 50ms CPU time on free tier, 30s on paid
• Memory limits: 128MB per request
• No long-running connections: Use Durable Objects for websockets

For session state, consider using Cloudflare KV or Durable Objects.

Testing Locally

Test your worker locally before deploying:

# Start local development server
npx wrangler dev

# Test the MCP endpoint
curl -X POST http://localhost:8787/mcp \
  -H "Content-Type: application/json" \
  -H "Accept: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": 1,
    "method": "tools/list"
  }'

# Test health check
curl http://localhost:8787/health

Monitoring and Logs

View logs and analytics in the Cloudflare dashboard:

# Stream logs in real-time
npx wrangler tail

# View metrics
npx wrangler metrics

Next Steps

• Serverless Deployments - Deploy to AWS Lambda, Google Cloud Functions
• Production Checklist - Security and monitoring best practices
• Custom Routes - Learn more about Hono routing