Skip to content
Latest versionv1.7.2

The TypeScript MCP framework

Build and ship MCP servers with the code you already know. Give AI applications access to your tools, data, and workflows. Write TypeScript functions; xmcp handles the protocol, validation, and transport.

Shell
npx create-xmcp-app@latest

Core concepts

Three primitives connect your application to an AI client. Explore each one in the docs.

Tools

Functions that let AI applications take action.

Resources

Data and context that AI applications can read.

Prompts

Reusable starting points for guided AI conversations.

The framework

Everything you need to set up fast, customize with ease, and connect your apps.

File System Routing

Automatically register tools, prompts, and resources with zero configuration.

Integrations

Secure access with Better Auth's integration and monetize with Polar.

Middlewares

Add authentication, logging, or custom logic to intercept and process requests and responses.

Extensible Configuration

Customize every aspect of your MCP server with flexible configuration options.

Multiple Transport Support

Build and deploy servers with HTTP or STDIO transport protocols out of the box.

Plug & play to your apps

Initialize xmcp directly in your existing Next.js or Express projects with one command.

Getting started

From a new project to a running server in four steps.

Create a new xmcp app
Shell
npx create-xmcp-app@latest

? What is your project named? my-server
? Select a template: Default (Standard MCP server)
? Select a package manager: pnpm
? Select the transport you want to use: HTTP (runs on a server)
? Select components to initialize: Tools, Prompts, Resources

Creating a new xmcp app in /Users/you/my-server.

? Ok to continue? Yes
✔ Your xmcp app is ready

Next steps:
  cd my-server
  pnpm dev

xmcp — ready.
Configure your environment
xmcp.config.ts
import { XmcpConfig } from "xmcp";

const config: XmcpConfig = {
  http: {
    port: 3000,
    host: "0.0.0.0",
    endpoint: "/mcp",
    cors: {
      origin: "*",
      credentials: true,
    },
  },
};

export default config;
Add tools, resources, and prompts
src/tools/get-weather.ts
import { z } from "zod";
import { type InferSchema } from "xmcp";

export const schema = {
  location: z.string().describe("City name"),
};

export const metadata = {
  name: "get-weather",
  description: "Get current weather for a location",
};

export default async function getWeather({ 
  location 
}: InferSchema<typeof schema>) {
  // Your implementation here
  return `Weather in ${location}: 72°F, Sunny`;
}
Deploy your server
Shell
vc deploy

Vercel CLI 37.0.0
Inspect: https://vercel.com/...
Production: https://my-server.vercel.app

Deployment Summary:
  • Environment: Production
  • Region: iad1
  • Build Time: 12s
  • Status: Ready

Your MCP server is live!

xmcp — ready.

From the blog

From TypeScript functions to MCP tools
Guides

From TypeScript functions to MCP tools

Reuse typed functions as MCP tools with experimental schema inference in xmcp. Generate input validation from TypeScript and JSDoc, and keep explicit schemas where you need them.

xmcp v1 is here
Changelog

xmcp v1 is here

xmcp v1 splits the compiler out of the runtime, moves to MCP revision 2026-07-28 through SDK v2, and keeps every existing tool, prompt and resource working unchanged.

When your agent has to pay
Engineering

When your agent has to pay

Charging for a tool is the easy half. The real shift is software that pays for tools on its own — how x402 lets agents settle USDC micropayments per call.

Integrating NestJS with xmcp
Guides

Integrating NestJS with xmcp

Add an MCP server to your NestJS application with tool discovery and a customizable module.

One framework to rule them all