Happy slice of mcp pizzaMCP.pizza

mcp-chatbot

MCP.Pizza Chef: 3choff

Built as a teaching example rather than a finished product, this small Python program runs in a terminal window, reads the same settings file format Claude Desktop uses, starts whichever MCP tools you list, and lets any OpenAI-compatible model call them. You bring your own model key. There have been no code changes since December 2024, so expect gaps against newer parts of the protocol, and treat it as a starting point to read and adapt rather than an app to depend on.

Unmaintained · no code changes since December 2024
Coding

Use This MCP client To

Chat with my MCP tools without installing a desktop app Try a model other than Claude with the same tools See in readable code how a chat app wires up tools Test a tool I am building from a plain terminal Reuse the settings file I already wrote for Claude Desktop

README

MCP Chatbot

This chatbot example demonstrates how to integrate the Model Context Protocol (MCP) into a simple CLI chatbot. The implementation showcases MCP's flexibility by supporting multiple tools through MCP servers and is compatible with any LLM provider that follows OpenAI API standards.

If you find this project helpful, don’t forget to ⭐ star the repository or buy me a ☕ coffee.

Key Features

  • LLM Provider Flexibility: Works with any LLM that follows OpenAI API standards (tested with Llama 3.2 90b on Groq and GPT-4o mini on GitHub Marketplace).
  • Dynamic Tool Integration: Tools are declared in the system prompt, ensuring maximum compatibility across different LLMs.
  • Server Configuration: Supports multiple MCP servers through a simple JSON configuration file like the Claude Desktop App.

Requirements

  • Python 3.10
  • python-dotenv
  • requests
  • mcp
  • uvicorn

Installation

mcp-chatbot FAQ

Is this still maintained?
No. The code has not changed since December 2024, well over a year ago. It is not archived, but nothing has been fixed or updated in that time.
Is this something I install into Claude or ChatGPT?
No, it is the other side of the pairing. It is the chat program itself, and the tools plug into it.
Do I need a key?
Yes. You supply a key for whichever model service you use, saved in a small settings file alongside the program.
Which models does it work with?
Anything that follows the OpenAI request format. The author tested it with Llama 3.2 90b on Groq and GPT-4o mini.
Can I use this to run something like a database or web browser tool from my terminal?
Yes. That is exactly the demonstration. You list the tools you want in the settings file and the program starts and offers them to the model.
How hard is the setup?
This one is for developers. You download the code, install Python packages, create a settings file, and run it by hand.