MCP TypeScript SDK
A TypeScript SDK for building MCP servers and clients, providing standardized context provision for LLMs through tools, resources, and prompts.
README
MCP TypeScript SDK
Overview
The Model Context Protocol (MCP) allows applications to provide context for LLMs in a standardized way, separating the concerns of providing context from the actual LLM interaction.
This repository contains the TypeScript SDK implementation of the MCP specification. It runs on Node.js, Bun, and Deno, and ships:
- MCP server libraries (tools/resources/prompts, Streamable HTTP, stdio, auth helpers)
- MCP client libraries (transports, high-level helpers, OAuth helpers)
- Optional middleware packages for specific runtimes/frameworks (Express, Fastify, Hono, Node.js HTTP)
- Runnable examples (under
examples/)
Packages
This monorepo publishes split packages:
@modelcontextprotocol/server: build MCP servers@modelcontextprotocol/client: build MCP clients
Tool and prompt schemas use Standard Schema — bring Zod v4, Valibot, ArkType, or any compatible library.
They are intentionally thin adapters: they should not introduce new MCP functionality or business logic. See packages/middleware/README.md for details.
@modelcontextprotocol/node: Node.js Streamable HTTP transport wrapper forIncomingMessage/ServerResponse@modelcontextprotocol/express: Express helpers (app defaults + Host header validation)@modelcontextprotocol/fastify: Fastify helpers (app defaults + Host header validation)@modelcontextprotocol/hono: Hono helpers (app defaults + JSON body parsing hook + Host header validation)
Installation
Server
npm install @modelcontextprotocol/server
# or
bun add @modelcontextprotocol/server
# or
deno add npm:@modelcontextprotocol/server
Client
npm install @modelcontextprotocol/client
# or
bun add @modelcontextprotocol/client
# or
deno add npm:@modelcontextprotocol/client
Optional middleware packages
The SDK also publishes optional “middleware” packages that help you wire MCP into a specific runtime or web framework (for example Express, Fastify, Hono, or Node.js http).
These packages are intentionally thin adapters and should not introduce additional MCP features or business logic. See packages/middleware/README.md for details.
# Node.js HTTP (IncomingMessage/ServerResponse) Streamable HTTP transport:
npm install @modelcontextprotocol/node
# Express integration:
npm install @modelcontextprotocol/express express
# Fastify integration:
npm install @modelcontextprotocol/fastify fastify
# Hono integration:
npm install @modelcontextprotocol/hono hono
Getting Started
Here is what an MCP server looks like. This minimal example exposes a single greet tool over stdio:
import { McpServer } from '@modelcontextprotocol/server';
import { StdioServerTransport } from '@modelcontextprotocol/server/stdio';
import * as z from 'zod/v4';
const server = new McpServer({ name: 'greeting-server', version: '1.0.0' });
server.registerTool(
'greet',
{
description: 'Greet someone by name',
inputSchema: z.object({ name: z.string() })
},
async ({ name }) => ({
content: [{ type: 'text', text: `Hello, ${name}!` }]
})
);
async function main() {
const transport = new StdioServerTransport();
await server.connect(transport);
}
main();
Ready to build something real? Follow the step-by-step tutorials:
- Build your first server — a stdio weather-alert server, from
npm initto a tool call - Build your first client — connect to that server, list its tools, and call them
For runnable, end-to-end examples beyond the tutorials, see:
examples/README.md— runnable, self-verifying client/server example pairs (one story per directory)
Documentation
- Build a server — your first MCP server, step by step
- Build a client — your first MCP client, step by step
- Documentation site — the full guides: tools, resources, prompts, serving over HTTP and stdio, clients, OAuth, and migration
- Troubleshooting — common errors and their fixes
- API reference
- MCP documentation
- MCP specification
Building docs locally
To work on the documentation site locally:
pnpm docs:api # Generate the API reference markdown (output: docs/api/)
pnpm docs:dev # Start the VitePress dev server for the V2 site
pnpm docs:build # Build the V2 site (output: docs/.vitepress/dist/)
pnpm docs:multi # Build the combined V1 + V2 site (output: tmp/docs-combined/)
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。