FRD/TDD MCP Server

An MCP (Model Context Protocol) server that generates Functional Requirements Documents (FRD) and Technical Design Documents (TDD) using AI.

Features

• 📋 Generate FRD - Create comprehensive functional requirements from project descriptions
• 🛠️ Generate TDD - Generate technical design documents from existing FRD
• 🤖 AI-Powered - Uses Google Gemini 2.5 Flash with automatic fallback (FREE, reliable!)
• 🔌 MCP Compatible - Works seamlessly with GitHub Copilot and other MCP clients
• ⚡ Resilient - Automatically tries alternative models if one is unavailable

Installation

Option 1: Clone and Build Locally

git clone https://github.com/yourusername/frd-tdd-mcp.git
cd frd-tdd-mcp
npm install
npm run build

Option 2: Install from npm (when published)

npm install -g frd-tdd-mcp

Configuration

Option 1: Anthropic Claude (RECOMMENDED)

Why Claude?

• 🏆 Best Quality - Superior comprehension and structure
• 🚀 200K Context - Handles large inputs easily
• 💰 Generous Limits - Better quota than free Gemini
• ⚡ Fast & Reliable - Consistent performance
• 🔄 Multiple Models - Auto-fallback across 4 Claude versions

Available Claude Models (tries in order):

1. claude-3-5-sonnet-20241022 - Latest Sonnet (Oct 2024) - Best balance

2. claude-3-5-sonnet-20240620 - Previous Sonnet (June 2024)

3. claude-3-opus-20240229 - Most capable model

4. claude-3-haiku-20240307 - Fastest, most affordable

5. Get API key: Anthropic Console

6. Add to MCP config (%APPDATA%\Code\User\mcp.json):

{
  "servers": {
    "frd-tdd": {
      "command": "node",
      "args": ["c:\\path\\to\\frd-tdd-mcp\\build\\index.js"],
      "env": {
        "ANTHROPIC_API_KEY": "your-anthropic-key-here"
      }
    }
  }
}

Option 2: Google Gemini (Free Fallback)

1. Get API key: Google AI Studio
2. Add to MCP config:

{
  "servers": {
    "frd-tdd": {
      "command": "node",
      "args": ["c:\\path\\to\\frd-tdd-mcp\\build\\index.js"],
      "env": {
        "GEMINI_API_KEY": "your-gemini-key-here"
      }
    }
  }
}

Option 3: Both (Best Reliability)

{
  "servers": {
    "frd-tdd": {
      "command": "node",
      "args": ["c:\\path\\to\\frd-tdd-mcp\\build\\index.js"],
      "env": {
        "ANTHROPIC_API_KEY": "your-anthropic-key-here",
        "GEMINI_API_KEY": "your-gemini-key-here"
      }
    }
  }
}

Priority: Anthropic → Gemini (automatic fallback if primary fails)

3. Restart VS Code

Usage

In GitHub Copilot Chat:

Generate FRD:

Generate an FRD for an e-commerce platform with shopping cart and payment processing

Generate TDD:

Generate a TDD from the current FRD

Available Tools

| Tool | Description | Parameters | |------|-------------|------------| | generate_frd | Generate Functional Requirements Document | description: Project description | | generate_tdd | Generate Technical Design Document | None (reads from FRD.md) |

Output

• FRD.md - Generated in project root with complete requirements
• TDD.md - Generated in project root with technical specifications

Development

# Install dependencies
npm install

# Run in development mode
npm start

# Build for production
npm run build

# Test the MCP server
npm run mcp

Project Structure

frd-tdd-mcp/
├── src/
│   ├── index.ts           # MCP server entry point
│   ├── interactive.ts     # Interactive CLI mode
│   ├── services/
│   │   └── aiServices.ts  # Groq AI integration
│   └── tools/
│       ├── generateFRD.ts # FRD generation logic
│       └── generateTDD.ts # TDD generation logic
├── build/                 # Compiled output
├── package.json
└── tsconfig.json

Requirements

• Node.js 18+
• Groq API key
• VS Code with GitHub Copilot

License

MIT

Author

Your Name

Contributing

Pull requests welcome! Please ensure tests pass before submitting.

Support

For issues and questions, please open an issue on GitHub.