MCP Server in JavaScript

A Model Context Protocol (MCP) server implementation using JavaScript and the official MCP SDK.

Overview

This project demonstrates how to build an MCP server that can provide tools and resources to AI assistants. The Model Context Protocol enables secure, standardized connections between AI applications and external data sources.

Features

This MCP server includes the following tools:

• azure-java-sdk-code-samples-mcp-server - Get readme.md for Azure Java SDK, which includes introduction, key concepts, and code samples

Prerequisites

• Node.js (version 18 or higher)
• npm or yarn package manager

Installation

1. Clone or download this project
2. Install dependencies:

npm install

Usage

Running the Server

Start the MCP server:

npm start

For development with debugging:

npm run dev

Or press F5 in VS Code to start debugging.

Connecting to AI Assistants

This MCP server can be connected to various AI assistants that support the Model Context Protocol:

1. Claude Desktop - See MCP_CONFIG.md for configuration details
2. Continue.dev - VS Code extension that supports MCP
3. Other MCP clients - Any client that supports stdio transport

VS Code

Sample mcp.json for connect locally.

{
  // 💡 Inputs are prompted on first server start, then stored securely by VS Code.
  "inputs": [
  ],
  "servers": {
    "azure-java-sdk-mcp": {
      "type": "stdio",
      "command": "node",
      "args": [
        "<path-to-index.js>"
      ]
    }
  }
}

One may need to "start" it once, to cache the tools.

Available VS Code Tasks

• Run MCP Server - Starts the server normally
• Debug MCP Server - Starts with Node.js inspector for debugging

Development

Project Structure

├── index.js                     # Main MCP server implementation
├── package.json                 # Project configuration
├── .github/
│   └── copilot-instructions.md  # Copilot customization
└── README.md                    # This file

Adding New Tools

To add a new tool:

1. Add the tool definition in the ListToolsRequestSchema handler
2. Implement the tool logic in the CallToolRequestSchema handler
3. Include proper input validation and error handling

Error Handling

The server uses McpError for proper error responses:

• ErrorCode.InvalidRequest - For invalid parameters
• ErrorCode.MethodNotFound - For unknown tools
• ErrorCode.InternalError - For unexpected errors

MCP Protocol

This server implements the Model Context Protocol specification. For more information about MCP:

• MCP Documentation
• MCP Specification

License

ISC License