Example MCP Server
A Node.js and TypeScript implementation that provides system information and health check tools for Claude Desktop. It serves as a boilerplate for building and integrating custom tools using the Model Context Protocol.
README
Example MCP Server
A complete Model Context Protocol (MCP) server implementation in Node.js + TypeScript that integrates with Claude Desktop as a local tool.
What is MCP?
The Model Context Protocol (MCP) is a standardized protocol that enables large language models (like Claude) to safely interact with external tools and data sources. MCP servers expose capabilities (tools) that Claude can discover and invoke in real-time during conversations.
Key Features of MCP:
- Tool Discovery: Claude can query available tools and their parameters
- Safe Execution: Structured JSON-RPC messages with validation
- Stdio Transport: For local integration with Claude Desktop
- Extensible: Easy to add new tools with standardized interfaces
Project Structure
example-mcp-server/
├── src/
│ ├── mcp/
│ │ └── server.ts # MCP server with stdio transport
│ ├── tools/
│ │ ├── index.ts # Tool registry & dispatcher
│ │ ├── ping.ts # Ping tool (health check)
│ │ └── system_info.ts # System info tool
│ └── types/
│ └── index.ts # Type definitions
├── package.json # Dependencies & scripts
├── tsconfig.json # TypeScript config
├── mcp.json # Claude Desktop manifest
└── README.md # This file
Getting Started
Prerequisites
- Node.js 18+ (LTS)
- npm or yarn
- Claude Desktop (for testing)
Installation
- Clone or navigate to the project directory:
cd example-mcp-server
- Install dependencies:
npm install
- Build the TypeScript code:
npm run build
This compiles all .ts files in src/ to JavaScript in the dist/ folder.
Running the MCP Server
npm start
The server will start listening on stdin/stdout, ready to receive JSON-RPC messages.
Testing the Server Locally
Using curl with echo (Windows PowerShell)
Test the tools/list endpoint:
$message = @{
jsonrpc = "2.0"
id = 1
method = "tools/list"
} | ConvertTo-Json -Compress
"$message" | node dist/mcp/server.js
Using a Node.js test script
Create test.js:
import { spawn } from "child_process";
const server = spawn("node", ["dist/mcp/server.js"]);
let output = "";
server.stdout.on("data", (data) => {
const line = data.toString().trim();
if (line) {
console.log("Response:", JSON.parse(line));
server.kill();
}
});
server.on("close", () => {
console.log("Server closed");
});
// Send initialize request
const initMsg = {
jsonrpc: "2.0",
id: 1,
method: "initialize",
params: {
protocolVersion: "2024-11-05",
capabilities: {},
clientInfo: { name: "test-client", version: "1.0.0" },
},
};
server.stdin.write(JSON.stringify(initMsg) + "\n");
// Wait a moment, then send tools/list
setTimeout(() => {
const listMsg = { jsonrpc: "2.0", id: 2, method: "tools/list" };
server.stdin.write(JSON.stringify(listMsg) + "\n");
}, 100);
Run with:
npm run build && node test.js
Available Tools
1. ping
A simple health check tool.
Request:
{
"jsonrpc": "2.0",
"id": 1,
"method": "tools/call",
"params": {
"name": "ping",
"arguments": {}
}
}
Response:
{
"jsonrpc": "2.0",
"id": 1,
"result": {
"output": "pong"
}
}
2. system_info
Returns system and platform information.
Request:
{
"jsonrpc": "2.0",
"id": 2,
"method": "tools/call",
"params": {
"name": "system_info",
"arguments": {}
}
}
Response:
{
"jsonrpc": "2.0",
"id": 2,
"result": {
"output": {
"platform": "win32",
"arch": "x64",
"nodeVersion": "v20.10.0",
"osType": "Windows_NT",
"osRelease": "10.0.22621",
"totalMemory": "16384 MB",
"freeMemory": "8192 MB",
"cpuCount": 8,
"uptime": "24 hours"
}
}
}
Registering with Claude Desktop
To make this MCP server available in Claude Desktop:
Windows
- Open
%APPDATA%\Claude\claude_desktop_config.json - Add this server configuration:
{
"mcpServers": {
"example-mcp-server": {
"command": "node",
"args": ["C:\\path\\to\\example-mcp-server\\dist\\mcp\\server.js"]
}
}
}
Replace C:\path\to\example-mcp-server with the actual full path to your project.
macOS
- Open
~/Library/Application Support/Claude/claude_desktop_config.json - Add the server configuration (use POSIX paths)
Linux
- Open
~/.config/Claude/claude_desktop_config.json - Add the server configuration
Example Configuration
{
"mcpServers": {
"example-mcp-server": {
"command": "node",
"args": ["C:\\Users\\YourUsername\\projects\\example-mcp-server\\dist\\mcp\\server.js"]
}
}
}
- Restart Claude Desktop for the changes to take effect
- In Claude, look for the tool icon or mention "use the ping tool" or "check system info"
Adding New Tools
To add a new tool:
- Create a tool file in
src/tools/(e.g.,src/tools/my_tool.ts):
import { ToolHandler } from "../types/index.js";
export const myToolHandler: ToolHandler = async (args) => {
// Implement your tool logic
return { result: "Your result here" };
};
export const myToolDefinition = {
name: "my_tool",
description: "Description of what your tool does",
inputSchema: {
type: "object",
properties: {
param1: { type: "string", description: "First parameter" },
param2: { type: "number", description: "Second parameter" },
},
required: ["param1"],
},
};
- Register it in
src/tools/index.ts:
import { myToolHandler, myToolDefinition } from "./my_tool.js";
const registry: ToolRegistry = {
// ... existing tools
my_tool: {
definition: myToolDefinition,
handler: myToolHandler,
},
};
- Rebuild and restart:
npm run build
npm start
MCP Protocol Overview
The server implements the MCP specification with support for:
- initialize: Handshake with client
- tools/list: Returns available tools and schemas
- tools/call: Executes a tool with provided arguments
All communication uses JSON-RPC 2.0 format with structured error handling.
Troubleshooting
Server won't start
- Ensure Node.js 18+ is installed:
node --version - Check TypeScript compiled correctly:
npm run build - Verify
dist/mcp/server.jsexists
Claude Desktop doesn't see the server
- Double-check the path in
claude_desktop_config.jsonis correct and absolute - Restart Claude Desktop after updating config
- Check server runs locally:
npm start
Tool calls fail
- Ensure tool name matches exactly (case-sensitive)
- Verify required parameters are provided
- Check error message in JSON-RPC response
Scripts
npm run build # Compile TypeScript to dist/
npm start # Run the MCP server
npm run dev # Build and run in one command
License
MIT
References
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
E2B
Using MCP to run code via e2b.