MCP Research Assistant

A powerful research assistant built using the Model Context Protocol (MCP) that helps users search, analyze, and organize academic papers from arXiv.

Features

• Paper Search: Search for academic papers on any topic using arXiv's API
• Paper Organization: Automatically organize papers by topic in a structured format
• Paper Information Extraction: Extract and store detailed information about papers including:
  • Title
  • Authors
  • Publication date
  • Summary
  • PDF URL
• Interactive Chat Interface: Chat with an AI assistant to help with your research queries
• Topic-based Organization: Papers are automatically organized into topic-based folders
• Comprehensive Research Analysis: Get summaries and analysis of research trends in your area of interest

Prerequisites

• Python 3.x (used 3.12.8)
• Node.js and npm (used v22.13.0)
• Anthropic API key

Installation

1. Clone the repository:

git clone https://github.com/ezequiroga/mcp-bases
cd mcp-bases

2. Install Python dependencies:

pip install -r requirements.txt

3. Copy the environment file and add your API key:

cp .env-example .env
# Edit .env and add your Anthropic API key

4. Update npm

npm install -g npm@latest

Project Structure

• research_server.py: Main server implementation for paper search and management
• mcp_chatbot.py: Interactive chatbot implementation
• server_config.json: Configuration for MCP servers
• papers/: Directory where paper information is stored (created automatically)

Usage

1. Start the research assistant:

python mcp_chatbot.py

2. Interact with the chatbot using natural language queries. For example:

   • "Search for papers about quantum computing"
   • "What are the latest papers on machine learning?"
   • "Summarize the research on artificial intelligence"

3. Finish the chatbot by typing quit.

Available Tools

The system provides several tools through the MCP interface:

• search_papers: Search for papers on a specific topic
• extract_info: Get detailed information about a specific paper
• get_available_folders: List all available topic folders
• get_topic_papers: Get detailed information about papers in a specific topic

Configuration

The server_config.json file configures three MCP servers:

• filesystem: For file system operations
• research: The main research server
• fetch: For fetching operations