trackmcp
Back to directory
haroldadmin

fastify-mcp

View on GitHub

A Fastify plugin to run Model Context Protocol (MCP) servers

21 stars TypeScriptAI & Machine Learning Updated Oct 27, 2025
aifastifyjavascriptmodel-context-protocol

Documentation

fastify-mcp

Integrate Model Context Protocol servers with your Fastify app.

Supports the Streamable HTTP transport as well as the legacy HTTP+SSE transport.

Usage

First, define your MCP server.

ts
function createServer() {
  const mcpServer = new McpServer({
    name: "...",
    version: "...",
  });

  mcpServer.tool("...");
  mcpServer.resource("...");

  return mcpServer.server;
}

Create a Fastify app and register the plugin.

ts
import { fastify } from "fastify";
import { streamableHttp } from "fastify-mcp";
import { StreamableHTTPServerTransport } from "@modelcontextprotocol/sdk/server/streamableHttp.js";

const app = fastify();

app.register(streamableHttp, {
  // Set to `true` if you want a stateful server
  stateful: false,
  mcpEndpoint: "/mcp",
  sessions: new Sessions()
  createServer,
});

app.listen({ port: 8080 });

See the examples directory for more detailed examples.

Installation

bash
# npm
npm install fastify-mcp

# yarn
yarn add fastify-mcp

Session Management

The official MCP TypeScript SDK does not support managing multiple sessions out of the box, and therefore it's the host server's responsibility to do so.

This package uses an in-memory mapping of each active session against its session ID to manage multiple sessions, as recommended by the MCP SDK examples.

Session Events

The `Sessions` class emits the following events:

  • `connected`: Emitted when a new session is added.
  • `terminated`: Emitted when a session is removed.
  • `error`: Emitted when an asynchronous event handler throws an error.
ts
const sessions = new Sessions();

sessions.on("connected", (sessionId) => {
  console.log(`Session ${sessionId} connected`);
});

sessions.on("terminated", (sessionId) => {
  console.log(`Session ${sessionId} terminated`);
});

Contributing

Please file an issue if you encounter a problem when using this package. Pull requests for new features or bug fixes are also welcome.

Frequently asked questions

What is fastify-mcp?

fastify-mcp is A Fastify plugin to run Model Context Protocol (MCP) servers

How do I install fastify-mcp?

Open the GitHub repository and follow its README. Most MCP servers are added to your client's MCP config, then called by your agent.

Is fastify-mcp open source?

Yes — it is hosted on GitHub at https://github.com/haroldadmin/fastify-mcp and has 21 stars.

Related MCP tools

ComposioHQcomposio

Composio equips your AI agents & LLMs with 100+ high-quality integrations via function calling for the Model Context Protocol. Enhance AI assistants with powerf

25,863 TypeScript
agentic-aiagentsai+14
yamadashyrepomix

📦 Repomix is a powerful tool that packs your entire repository into a single, AI-friendly file. Perfect for when you need to feed your codebase to Large Lan...

19,761 TypeScript
aianthropicartificial-intelligence+17
mastra-aimastra

The TypeScript AI agent framework. ⚡ Assistants, RAG, observability. Supports any LLM: GPT-4, Claude, Gemini, Llama. Built for the Model Context Protocol to enh

18,004 TypeScript
agentsaichatbots+10
mcp-usemcp-use

Mcp-use is the easiest way to interact with mcp servers with custom agents TypeScript-based implementation. Trusted by 8100+ developers.

8,149 TypeScript
agentagentsai+6
zcaceresmarkdownify-mcp

A Model Context Protocol server for converting almost anything to Markdown TypeScript-based implementation. Trusted by 2200+ developers.

2,236 TypeScript
aianthropicanthropic-ai+6
danny-avilalibrechat

Enhanced ChatGPT Clone: Features Agents, MCP, DeepSeek, Anthropic, AWS, OpenAI, Responses API, Azure, Groq, o1, GPT-5, Mistral, OpenRouter, Vertex AI, Gemini...

31,284 TypeScript
aianthropicartifacts+17

Run your own MCP server? See who uses it and what to fix.

Measure it with TrackMCP