Project Overview

🚀 MCP Server + OpenAI Agent

This project integrates an MCP server with an OpenAI agent, enabling automated tasks such as web searches and file operations to enhance efficiency.

🚀 Quick Start

📦 Installation

1. Create and activate a virtual environment:

   • Run the provided setup.sh script to set up a virtual environment with Python 3.13.

     ./setup.sh
     

   • This script will create a virtual environment and install all the required Python packages according to requirements.txt.

2. Install project dependencies:

   • Ensure you have installed the external tools and services required to run the selected MCP servers:
     • Lynx: Required for the lynx server, which interfaces with the Lynx terminal web browser tool.
     • Docker: Required for the puppeteer and github servers to enable Chrome control and GitHub interaction.
     • Node.js: Required for the fs server to use the npx command.

💻 Usage Examples

Running agent.py

Basic Execution

• You can run the agent with all configured servers:

  python agent.py
  

• You can choose the servers the agent can access:

  # Allow the agent to access the file system (for local file browsing) and Lynx (for web searches)
  python agent.py --servers fs lynx
  

  In my experience, this can improve performance as the agent is less likely to choose the wrong tool.

• Debug Mode:

  • Use the --debug flag to initialize the servers without executing the agent. This is very useful for ensuring the servers are configured correctly.

    python agent.py --debug
    

Configuration

Servers are configured in servers.yaml, and each server entry includes:

• Name: The identifier of the server.
• Description: A brief description of the server's functionality.
• Commands and Parameters: The command-line instructions to start the server.
• Environment Variables: Any additional environmental configurations required to initialize the server. You can usually find these configuration files in the GitHub repository.

Custom Servers

You can develop your own MCP servers in the mcp - servers directory.

What Can This Project Do?

Link Search

If you're trying to set up QEMU, you can use agent.py for initial research:

"Find links related to QEMU setup" | python agent.py --servers lynx

How to Get Started?

Follow these steps to get started:

1. Run the setup.sh script to create a virtual environment and install dependencies.
2. Install the required external tools: Lynx, Docker, and Node.js.
3. Configure the servers.yaml file, adding or modifying server entries as needed.
4. Run the agent and test its functionality.