MCP Registry

A local application that aggregates Model Context Protocol (MCP) servers into a searchable directory with reputation scoring.

⚠️ Note: This repository is a live demo built to showcase the multi-agent AI development methodology.
While it is a functional product, its primary purpose is to demonstrate how you can use AI agents to build software — and even run entire workflows and teams — through a structured, repeatable prompt-based process.

The methodology behind this project is detailed in our upcoming blog post (link coming soon). It explores how roles like Product Manager, Architect, Designer, Dev Lead, and Engineers can be executed by AI agents, all collaborating through clear interfaces and shared outputs.

🗂 To explore the full process and see exactly how the app was built, check out the outputs generated by the different agents in the following directories:
product/, architecture/, design/, tasks/, and demo-methodology-files/.

This is a learning repository and we recognize it needs improvements in various areas. There are known vulnerabilities and test coverage limitations. In future posts, we'll demonstrate how to extend this project with proper CI/CD pipelines, Secure Software Development Lifecycle (SSDLC) practices, comprehensive testing, and other production-ready enhancements.

Project Overview

MCP Registry serves as a centralized directory for Model Context Protocol (MCP) servers. The system aggregates data from multiple sources, calculates reputation scores, and provides a searchable interface for users to discover and evaluate MCP servers.

Features

• Server Discovery: Find and browse MCP servers from various sources
• Reputation Scoring: Evaluate servers based on calculated reputation scores
• Search & Filter: Easily search and filter servers based on various criteria
• Local Database: All data is stored locally for offline access

Known Issues

Glama API Integration

There is a known issue with the Glama API integration for fetching MCP servers:

• The current implementation attempts to fetch data from the Glama.ai API endpoint, but there are connectivity problems with this external service.
• This affects the "Load Latest Servers" functionality on the homepage.
• Future versions will implement additional data sources and more robust error handling for external API dependencies.
• For now, the application works with locally submitted and mock data.

Technology Stack

• Frontend: React, TypeScript, Material-UI
• Backend: Node.js, Express, TypeScript
• Database: SQLite with Prisma ORM
• Application: Electron for desktop packaging

Getting Started

Prerequisites

• Node.js (v16+)
• npm or yarn

Installation

1. Clone the repository

   git clone https://github.com/yourusername/mcp-registry.git
   cd mcp-registry
   

2. Install dependencies

   npm install
   

3. Initialize Prisma

   npx prisma generate
   

Development

1. Start the backend server

   node server.js
   

   This will start the backend server on port 3001.

2. In a separate terminal, start the frontend development server

   npm run dev
   

   This will start the React application on port 5173.

3. Access the application in your browser at http://localhost:5173

Alternative Development Workflow

You can also use the combined script to start both servers:

npm run start-dev

Project Structure

mcp-registry/
├── src/
│   ├── main/                  # Electron main process
│   ├── renderer/              # Frontend (React)
│   │   ├── components/        # Reusable UI components
│   │   ├── pages/             # Page components
│   │   ├── services/          # Frontend services
│   │   ├── hooks/             # Custom React hooks
│   │   ├── types/             # TypeScript type definitions
│   │   ├── utils/             # Utility functions
│   ├── shared/                # Shared code between main and renderer
│   └── server/                # Backend (Express)
│       ├── controllers/       # API controllers
│       ├── models/            # Data models
│       ├── services/          # Backend services
│       │   ├── collection/    # Data collection service
│       │   ├── reputation/    # Reputation scoring engine
│       │   └── search/        # Search service
│       ├── utils/             # Utility functions
├── prisma/                    # Prisma schema and migrations
├── public/                    # Static assets
└── tests/                     # Test files

Available Scripts

• npm run dev - Start development mode
• npm run build - Build the application
• npm start - Start the built application
• npm test - Run tests
• npm run lint - Run ESLint
• npm run format - Format code with Prettier

License

This project is licensed under the MIT License - see the LICENSE file for details.