vm799/REG_MCP_SERVER

REG_MCP_SERVER 🤖📚

An intelligent MCP (Model Context Protocol) server that connects Claude AI with arXiv research papers. Search, analyze, and chat about academic papers with AI assistance.

🚨 Quick Fix for Access Token Error

Getting an "access token" error? This happens because the Anthropic API key is missing.

2-Minute Solution:

1. Get API Key 🔑

   • Visit console.anthropic.com
   • Sign up/login → API Keys → Create new key
   • Copy your key (starts with sk-ant-api03-...)

2. Configure API Key ⚙️

   # Edit the .env file
   nano .env
   
   # Replace the placeholder with your key:
   ANTHROPIC_API_KEY=sk-ant-api03-your-actual-key-here
   

3. Test & Run 🚀

   python test_setup.py  # Verify everything works
   python mcp_chatbot.py  # Start chatting!
   

✨ Features

• 🔍 Smart Paper Search - Find relevant research papers on arXiv by topic
• 📄 Paper Analysis - Extract detailed information (authors, summary, citations)
• 💬 AI-Powered Chat - Discuss papers with Claude AI
• 💾 Auto Storage - Automatically saves paper metadata locally
• 🏗️ MCP Architecture - Built on Model Context Protocol for seamless AI integration

📦 Installation & Setup

Prerequisites

• Python 3.12+
• uv package manager (recommended)
• Anthropic API key

Step 1: Clone & Install

# Clone the repository
git clone https://github.com/vm799/REG_MCP_SERVER.git
cd REG_MCP_SERVER

# Install dependencies with uv (recommended)
uv sync

# OR install with pip
pip install anthropic arxiv mcp nest-asyncio python-dotenv

Step 2: Configure API Key

# Copy environment template
cp .env.example .env

# Edit with your API key
nano .env

Add your Anthropic API key:

ANTHROPIC_API_KEY=sk-ant-api03-your-actual-key-here

Step 3: Test Setup

# Verify everything is configured correctly
python test_setup.py

You should see:

🧪 Testing REG_MCP_SERVER setup...
==================================================
✅ All required modules imported successfully
✅ Papers directory accessible
✅ Research server module imported successfully
✅ MCP tools (search_papers, extract_info) are available
✅ Environment file configured correctly
==================================================
🎉 All tests passed! Your MCP server is ready to use.

🚀 Usage

Start the Interactive Chatbot

python mcp_chatbot.py

Example Conversations

Search for Papers:

Query: Find papers about machine learning in healthcare
🔍 Calling tool search_papers with args {'topic': 'machine learning healthcare', 'max_results': 5}
Found 5 papers on machine learning in healthcare...

Get Paper Details:

Query: Tell me more about paper 2301.00001
📄 Here's detailed information about the paper including authors, summary, and key findings...

Ask Research Questions:

Query: What are the main challenges in applying ML to medical diagnosis?
💬 Based on the papers I found, here are the key challenges...

🏗️ Architecture

REG_MCP_SERVER/
├── mcp_chatbot.py      # Main chatbot interface
├── research_server.py  # MCP server with research tools
├── main.py            # Entry point
├── test_setup.py      # Setup verification
├── papers/            # Local paper storage
├── .env              # API key configuration
└── README.md         # This file

Core Components

1. MCP Chatbot (mcp_chatbot.py)

   • Handles user interactions
   • Manages conversation flow with Claude
   • Processes tool calls and responses

2. Research Server (research_server.py)

   • Provides MCP tools for paper search and extraction
   • Interfaces with arXiv API
   • Manages local paper storage

3. Tools Available:

   • search_papers(topic, max_results) - Search arXiv by topic
   • extract_info(paper_id) - Get detailed paper information

🛠️ Advanced Configuration

Custom Paper Storage

# Modify PAPER_DIR in research_server.py
PAPER_DIR = "custom_papers_directory"

Search Parameters

# Customize search behavior
search = arxiv.Search(
    query=topic,
    max_results=max_results,
    sort_by=arxiv.SortCriterion.Relevance,  # or SubmittedDate, LastUpdatedDate
    sort_order=arxiv.SortOrder.Descending
)

Claude Model Configuration

# Modify model in mcp_chatbot.py
model='claude-3-7-sonnet-20250219'  # Current model
# model='claude-3-haiku-20240307'    # Faster, cheaper
# model='claude-3-opus-20240229'     # Most capable

🔧 Troubleshooting

Common Issues

❌ "anthropic.AuthenticationError"

• Cause: Missing or invalid API key
• Fix: Check your .env file has the correct ANTHROPIC_API_KEY

❌ "ImportError: No module named 'anthropic'"

• Cause: Dependencies not installed
• Fix: Run uv sync or pip install -r requirements.txt

❌ "Permission denied writing to papers/"

• Cause: Insufficient write permissions
• Fix: chmod 755 papers/ or run with appropriate permissions

❌ "Connection timeout to arXiv"

• Cause: Network connectivity issues
• Fix: Check internet connection, try again later

Debug Mode

# Run with verbose output
python -v mcp_chatbot.py

# Check server logs
python research_server.py

Reset Paper Storage

# Clear all cached papers
rm -rf papers/
mkdir papers/

🤝 Contributing

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Make your changes
4. Add tests if applicable
5. Commit your changes (git commit -m 'Add amazing feature')
6. Push to the branch (git push origin feature/amazing-feature)
7. Open a Pull Request

Development Setup

# Install development dependencies
uv add --dev pytest black flake8 mypy

# Run tests
python -m pytest

# Format code
black *.py

# Check typing
mypy *.py

📚 API Reference

MCP Tools

search_papers(topic: str, max_results: int = 5) -> List[str]

Search for papers on arXiv based on a topic.

Parameters:

• topic: The research topic to search for
• max_results: Maximum number of papers to return (default: 5)

Returns:

• List of paper IDs found in the search

Example:

paper_ids = search_papers("neural networks", max_results=3)
# Returns: ['2301.00001', '2301.00002', '2301.00003']

extract_info(paper_id: str) -> str

Get detailed information about a specific paper.

Parameters:

• paper_id: The arXiv paper ID (e.g., "2301.00001")

Returns:

• JSON string with paper information or error message

Example:

info = extract_info("2301.00001")
# Returns detailed paper metadata as JSON

🔗 Useful Links

• Anthropic Console - Get API keys
• arXiv API Documentation - arXiv search details
• MCP Documentation - Model Context Protocol
• Claude API Docs - Anthropic API reference

📝 License

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

⭐ Support

If this project helped you, please give it a star! ⭐

Questions? Open an issue or start a discussion.

Made with ❤️ for researchers and AI enthusiasts