snowan/chatbot-mcp-server

Chatbot MCP Server

Overview

This repository implements an MCP-compatible research assistant chatbot system. It consists of:

• An MCP server that exposes research tools for searching and extracting arXiv paper information.
• An MCP client chatbot that interacts with the server and uses an LLM (Anthropic Claude) to process user queries and call tools.
• Jupyter notebooks that provide educational lessons, demos, and development workflows.

About This Project

This project follows the DeepLearning.AI course on understanding MCP (Model Context Protocol) and building a chatbot server as a demo. For more details and learning resources, see:

MCP: Build Rich-Context AI Apps with Anthropic (DeepLearning.AI Short Course)

Main Components

• research_server.py:
  The MCP server, exposing two tools:

  • search_papers(topic, max_results): Search arXiv and save paper metadata.
  • extract_info(paper_id): Retrieve metadata for a specific paper.

• mcp_chatbot.py:
  The MCP client chatbot, which:

  • Connects to the server via stdio.
  • Uses Anthropic Claude for LLM-based chat.
  • Handles tool-calling and interactive chat loop.

• Jupyter Notebooks:

  • Provide step-by-step lessons on building the server, client, and integrating with MCP.
  • Include code examples, explanations, and environment setup instructions.

Dependencies

• Python 3.13+
• anthropic (LLM API)
• arxiv (arXiv API)
• mcp (MCP protocol client/server)
• nest-asyncio, python-dotenv
• (See pyproject.toml for details)

Installation

1. Clone the repository:

   git clone <repo-url>
   cd chatbot-mcp-server
   

2. Set up a virtual environment:

   python3 -m venv .venv
   source .venv/bin/activate
   

3. Install dependencies:

   pip install -r requirements.txt
   # or, if using uv:
   uv pip install -r requirements.txt
   

4. Set up environment variables:

   • Create a .env file with your Anthropic API key and any other required variables.

Usage

Running the MCP Server

uv run research_server.py

• This starts the server and exposes the research tools via MCP protocol.

Running the Chatbot Client

python mcp_chatbot.py

• This launches the interactive chatbot, which connects to the server and allows you to query research papers.

Example Workflow

1. Start the server (uv run research_server.py).
2. In another terminal, start the chatbot (python mcp_chatbot.py).
3. Type a research query (e.g., "Find recent papers on quantum computing").
4. The chatbot will use the LLM to process your query, call the appropriate tool, and return results.

Notebooks and Demos

• chatbot-mcp-client.ipynb:
  Lesson on building the MCP client chatbot.
• chatbot-mcp-server-demo.ipynb:
  Lesson on building the MCP server.
• chatbot-demo.ipynb:
  Standalone chatbot example.
• chatbot-mcp-server.ipynb:
  Full example integrating server and client.

These notebooks are recommended for step-by-step learning and experimentation.

Project Structure

chatbot-mcp-server/
├── mcp_chatbot.py
├── research_server.py
├── main.py
├── papers/
├── .venv/ or venv/
├── pyproject.toml
├── uv.lock
├── .python-version
├── README.md
├── *.ipynb (notebooks)
└── .gitignore

Contributing

Contributions are welcome! Please open issues or submit pull requests for improvements, bug fixes, or new features.