MCP Server Boilerplate
A boilerplate server implementation for the Model Context Protocol (MCP), built with TypeScript and Express.
Table of Contents
- Overview
- Project Structure
- Getting Started
- Running the Server
- Resources
- Tools
- Prompts
- Extending the Server
- Testing and Debugging
- License
Overview
This project implements a server that follows the Model Context Protocol (MCP), which allows applications to provide context for LLMs in a standardized way. It includes:
- A fully configured MCP server with HTTP and stdio transport options
- Sample resources, tools, and prompts to demonstrate key functionality
- TypeScript support for type safety and better developer experience
- Express integration for the HTTP transport layer
Project Structure
mcp-server-boilerplate/
├── .env # Environment variables
├── .env.example # Example environment variables
├── .gitignore # Git ignore file
├── package.json # Project dependencies and scripts
├── tsconfig.json # TypeScript configuration
├── src/
│ ├── index.ts # Main HTTP server entry point
│ ├── stdio.ts # Stdio server entry point
│ ├── resources/ # MCP resources
│ │ ├── index.ts # Resource registration
│ │ ├── infoResource.ts # Static info resource
│ │ └── greetingResource.ts # Dynamic greeting resource
│ ├── tools/ # MCP tools
│ │ ├── index.ts # Tool registration
│ │ ├── calculatorTool.ts # Sample calculator tool
│ │ └── timestampTool.ts # Sample timestamp tool
│ └── prompts/ # MCP prompts
│ ├── index.ts # Prompt registration
│ ├── greetingPrompt.ts # Sample greeting prompt
│ └── analyzeDataPrompt.ts # Sample data analysis prompt
└── dist/ # Compiled JavaScript output
Getting Started
Prerequisites
- Node.js (v18 or later)
- npm or yarn
Installation
Clone the repository and install dependencies:
git clone https://github.com/yourusername/mcp-server-boilerplate.git
cd mcp-server-boilerplate
npm install
Environment Variables
Copy the example environment file and modify as needed:
cp .env.example .env
Available environment variables:
PORT: The port for the HTTP server (default: 3000)NODE_ENV: Environment mode (development, production)- OAuth settings (if needed)
Running the Server
HTTP Server
Build and start the HTTP server:
npm run build
npm start
For development with auto-restart:
npm run dev
The server will be available at http://localhost:3000/mcp (or the port specified in your .env file).
Stdio Mode
To run the server in stdio mode (for command-line tools):
npm run start:stdio
For development with auto-restart:
npm run dev:stdio
Resources
The boilerplate includes these example resources:
-
Static Info Resource:
info://server- Provides basic information about the server
-
Dynamic Greeting Resource:
greeting://{name}- Generates a personalized greeting with the provided name parameter
To access resources:
- Through the MCP protocol
- Using an MCP client library
Tools
The boilerplate includes these example tools:
-
Calculator: Performs basic arithmetic operations
- Parameters:
operation: Operation to perform (add, subtract, multiply, divide)a: First numberb: Second number
- Parameters:
-
Timestamp: Provides the current time in various formats
- Parameters:
format: Output format (iso, unix, readable)
- Parameters:
Prompts
The boilerplate includes these example prompts:
-
Greeting: Creates a personalized greeting prompt
- Parameters:
name: Name to greetformal: Whether to use formal greeting style (optional)
- Parameters:
-
Analyze Data: Creates a prompt for data analysis
- Parameters:
data: The data to analyzeformat: Data format (json, csv, text)instructions: Additional analysis instructions (optional)
- Parameters:
Extending the Server
Adding Resources
To add a new resource:
- Create a new file in
src/resources/(e.g.,myResource.ts) - Implement your resource handler
- Register it in
src/resources/index.ts
Example:
// myResource.ts
import { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js';
export function myResource(server: McpServer): void {
server.resource('my-resource', 'my-resource://path', async uri => ({
contents: [
{
uri: uri.href,
text: 'My resource content',
},
],
}));
}
// Then add to resources/index.ts
import { myResource } from './myResource.js';
export function registerResources(server: McpServer): void {
// ...existing resources
myResource(server);
}
Adding Tools
To add a new tool:
- Create a new file in
src/tools/(e.g.,myTool.ts) - Implement your tool handler
- Register it in
src/tools/index.ts
Example:
// myTool.ts
import { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js';
import { z } from 'zod';
export function myTool(server: McpServer): void {
server.tool('my-tool', { param: z.string() }, async ({ param }) => ({
content: [
{
type: 'text',
text: `Processed: ${param}`,
},
],
}));
}
// Then add to tools/index.ts
import { myTool } from './myTool.js';
export function registerTools(server: McpServer): void {
// ...existing tools
myTool(server);
}
Adding Prompts
To add a new prompt:
- Create a new file in
src/prompts/(e.g.,myPrompt.ts) - Implement your prompt handler
- Register it in
src/prompts/index.ts
Example:
// myPrompt.ts
import { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js';
import { z } from 'zod';
export function myPrompt(server: McpServer): void {
server.prompt('my-prompt', { topic: z.string() }, ({ topic }) => ({
messages: [
{
role: 'user',
content: {
type: 'text',
text: `Please explain ${topic} in simple terms.`,
},
},
],
}));
}
// Then add to prompts/index.ts
import { myPrompt } from './myPrompt.js';
export function registerPrompts(server: McpServer): void {
// ...existing prompts
myPrompt(server);
}
Testing and Debugging
To test your MCP server, you can use:
- The MCP Inspector tool
- MCP client libraries
- Direct HTTP requests (for debugging)
License
This project is licensed under the MIT License - see the LICENSE file for details.