MCP Research Assistant 🧠

A comprehensive Model Context Protocol (MCP) setup that provides powerful tools for research, file management, and web content fetching. This project integrates multiple MCP servers to enhance your AI assistant capabilities.

✨ Features

• 📚 Research Tool: Search and manage academic papers from arXiv
• 📁 Filesystem Tool: Browse, read, and manage project files
• 🌐 Fetch Tool: Retrieve content from websites and APIs
• 🤖 Multi-LLM Support: Works with Claude, Gemini, and other AI models
• 💾 Local Storage: Automatically saves research data organized by topics

🛠️ Prerequisites

• Python 3.13 or higher
• uv package manager (recommended) or pip
• API keys for your chosen LLM providers
• Claude Desktop (for MCP integration)

💻 Quick Start

1. Clone and Setup

git clone <your-repo-url>
cd mcp_project

2. Install Dependencies

# Install uv if you haven't already
curl -LsSf https://astral.sh/uv/install.sh | sh

# Create virtual environment and install dependencies
uv sync

3. Configure Environment Variables

Create a .env file in your project root:

# Choose one or both depending on your needs
ANTHROPIC_API_KEY=your_anthropic_api_key_here
GOOGLE_API_KEY=your_google_api_key_here

4. Configure Claude Desktop

Create or update your Claude Desktop configuration file:

Location: ~/Library/Application Support/Claude/claude_desktop_config.json (macOS)

{
  "mcpServers": {
    "filesystem": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-filesystem",
        "."
      ],
      "cwd": "/path/to/your/mcp_project"
    },
    "research": {
      "command": "/path/to/your/mcp_project/.venv/bin/python",
      "args": [
        "/path/to/your/mcp_project/research_server.py"
      ],
      "cwd": "/path/to/your/mcp_project"
    },
    "fetch": {
      "command": "/path/to/your/.local/bin/uvx",
      "args": ["mcp-server-fetch"],
      "cwd": "/path/to/your/mcp_project"
    }
  }
}

Important: Replace /path/to/your/mcp_project with your actual project path.

5. Restart Claude Desktop

Restart Claude Desktop completely to load the new configuration.

🎯 How to Use

Research Tool 🔬

Search for Papers:

Search for 5 papers about machine learning

Get Paper Details:

Show me information about paper ID 1234.5678

Browse Saved Papers:

What papers do I have saved on physics?

Filesystem Tool 📁

Browse Files:

List all files in my project directory

Read Files:

Show me the contents of research_server.py

Create Files:

Create a new Python script for data analysis

Fetch Tool 🌐

Get Web Content:

Fetch the latest Python documentation

API Calls:

Get current weather data from an API

📋 Available Tools

Research Server Tools

Filesystem Server Tools

Fetch Server Tools

📁 Project Structure

mcp_project/
├── research_server.py          # Main research MCP server
├── mcp_chatbot_L7.py          # Chatbot with LLM integration
├── pyproject.toml             # Project configuration
├── requirements.txt           # Python dependencies
├── uv.lock                   # Dependency lock file
├── papers/                   # Research data storage
│   └── [topic_name]/         # Organized by topic
│       └── papers_info.json  # Paper metadata
├── .env                      # Environment variables
└── README.md                 # This file

🔧 Configuration Details

Research Server Configuration

The research server automatically:

• Creates topic-based directories in papers/
• Saves paper metadata as JSON files
• Provides search and retrieval functions
• Integrates with arXiv API

Filesystem Server Configuration

The filesystem server:

• Operates within your project directory
• Provides full file management capabilities
• Uses relative paths for portability

Fetch Server Configuration

The fetch server:

• Handles web requests and API calls
• Supports custom user agents
• Can ignore robots.txt restrictions

🔄 Updating Dependencies

# Add new dependencies
uv add package_name

# Update requirements.txt
uv pip freeze > requirements.txt

# Sync all dependencies
uv sync

📝 Development

Adding New Tools

1. Edit research_server.py to add new functions
2. Use the @mcp.tool() decorator
3. Test with MCP Inspector
4. Update documentation

Customizing LLM Behavior

1. Edit mcp_chatbot_L7.py
2. Modify tool descriptions and parameters
3. Add custom prompts and resources