MCP Research Server

A demonstration project showcasing how to create a custom MCP (Model Control Protocol) server using the FastMCP framework. This implementation serves as a practical example of building an MCP server that integrates with arXiv to search, retrieve, and manage academic papers, providing a structured interface for AI-assisted research. The project leverages FastMCP's capabilities to create custom tools, resources, and prompts, making it an excellent reference for developing your own MCP servers.

🚀 Features

• Paper Search: Search for academic papers on arXiv by topic
• Persistent Storage: Automatically saves paper metadata for future reference
• Topic Organization: Organizes papers by research topics
• RESTful API: Exposes endpoints through FastMCP for easy integration
• AI Integration: Includes prompts for Claude AI to analyze research papers

🛠️ Prerequisites

• Python 3.11.11 (recommended)
• pip (Python package manager)
• Virtual environment (recommended)

Note: This project is developed and tested with Python 3.11.11. While it may work with other Python 3.7+ versions, using 3.11.11 is recommended for compatibility.

🚀 Installation

1. Clone the repository:

   git clone https://github.com/teddytesfa/research-mcp-server.git
   cd mcp_project
   

2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`
   

3. Install the required packages:

   pip install -r requirements.txt
   

🏃‍♂️ Getting Started

1. Start the MCP server:

   python research_server.py
   

2. The server will start on http://localhost:8001 by default.

🛠️ API Endpoints

Tools

• search_papers(topic: str, max_results: int = 5)

  • Search for papers on arXiv
  • Returns: List of paper IDs

• extract_info(paper_id: str)

  • Get detailed information about a specific paper
  • Returns: Paper metadata as JSON

Resources

• papers://folders

  • List all available topic folders

• papers://{topic}

  • Get papers for a specific topic

🤖 AI Integration

The server includes a prompt template for Claude AI to analyze research papers. Use the generate_search_prompt function to get started with AI-assisted research.

📁 Project Structure

mcp_project/
├── papers/                  # Directory for storing paper metadata
│   └── {topic}/             # Topic-specific directories
│       └── papers_info.json  # Paper metadata
├── research_server.py        # Main server implementation
└── README.md                # This file

📝 Example Usage

1. Search for papers on a topic:

   # Example search for papers about "machine learning"
   paper_ids = search_papers("machine learning", max_results=3)
   

2. Get information about a specific paper:

   paper_info = extract_info("2103.00001")
   print(paper_info)
   

📚 Learning Resources

This project is an implementation to demonstrate how to create your own MCP (Model Control Protocol) server. It showcases:

• Creating custom tools with @mcp.tool()
• Defining resources with @mcp.resource()
• Generating AI prompts with @mcp.prompt()
• Managing server state and persistence
• Integrating with external APIs (arXiv)

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.