xmcp Application

This project was created with create-xmcp-app.

Getting Started

First, run the development server:

npm run dev
# or
yarn dev
# or
pnpm dev

This will start the MCP server with the selected transport method.

Related MCP server: MCP Inspector Server

Project Structure

This project uses the structured approach where tools, prompts, and resources are automatically discovered from their respective directories:

• src/tools - Tool definitions

• src/prompts - Prompt templates

• src/resources - Resource handlers

Tools

Each tool is defined in its own file with the following structure:

import { z } from "zod";
import { type InferSchema, type ToolMetadata } from "xmcp";

export const schema = {
  name: z.string().describe("The name of the user to greet"),
};

export const metadata: ToolMetadata = {
  name: "greet",
  description: "Greet the user",
  annotations: {
    title: "Greet the user",
    readOnlyHint: true,
    destructiveHint: false,
    idempotentHint: true,
  },
};

export default function greet({ name }: InferSchema<typeof schema>) {
  return `Hello, ${name}!`;
}

Prompts

Prompts are template definitions for AI interactions:

import { z } from "zod";
import { type InferSchema, type PromptMetadata } from "xmcp";

export const schema = {
  code: z.string().describe("The code to review"),
};

export const metadata: PromptMetadata = {
  name: "review-code",
  title: "Review Code",
  description: "Review code for best practices and potential issues",
  role: "user",
};

export default function reviewCode({ code }: InferSchema<typeof schema>) {
  return `Please review this code: ${code}`;
}

Resources

Resources provide data or content with URI-based access:

import { z } from "zod";
import { type ResourceMetadata, type InferSchema } from "xmcp";

export const schema = {
  userId: z.string().describe("The ID of the user"),
};

export const metadata: ResourceMetadata = {
  name: "user-profile",
  title: "User Profile",
  description: "User profile information",
};

export default function handler({ userId }: InferSchema<typeof schema>) {
  return `Profile data for user ${userId}`;
}

Adding New Components

Adding New Tools

To add a new tool:

1. Create a new .ts file in the src/tools directory

2. Export a schema object defining the tool parameters using Zod

3. Export a metadata object with tool information

4. Export a default function that implements the tool logic

Adding New Prompts

To add a new prompt:

1. Create a new .ts file in the src/prompts directory

2. Export a schema object defining the prompt parameters using Zod

3. Export a metadata object with prompt information and role

4. Export a default function that returns the prompt text

Adding New Resources

To add a new resource:

1. Create a new .ts file in the src/resources directory

2. Use folder structure to define the URI (e.g., (users)/[userId]/profile.ts → users://{userId}/profile)

3. Export a schema object for dynamic parameters (optional for static resources)

4. Export a metadata object with resource information

5. Export a default function that returns the resource content

Building for Production

To build your project for production:

npm run build
# or
yarn build
# or
pnpm build

This will compile your TypeScript code and output it to the dist directory.

Running the Server

You can run the server for the transport built with:

• HTTP: node dist/http.js

• STDIO: node dist/stdio.js

Given the selected transport method, you will have a custom start script added to the package.json file.

For HTTP:

npm run start-http
# or
yarn start-http
# or
pnpm start-http

For STDIO:

npm run start-stdio
# or
yarn start-stdio
# or
pnpm start-stdio

Learn More

• xmcp Documentation