MyMCP Prompt

What is MyMCP Prompt?

MyMCP Prompt is a tool for generating Model Context Protocol (MCP) servers from natural language descriptions. It uses the Google Gemini API to convert user descriptions into functional Python MCP servers with corresponding JSON configurations. The application consists of a Flask backend and a React frontend built with Vite.

How to use MyMCP Prompt?

Clone the repository, set up the Python virtual environment and install dependencies, then install Node.js dependencies for the frontend. Set the GEMINI_API_KEY environment variable with a valid Google Gemini API key. Start the Flask backend with flask run and the frontend dev server with npm run dev. Open http://localhost:5173, enter a server description, and click "Place Your Order" to generate the server code and configuration.

Key features of MyMCP Prompt

• Generates MCP server code from natural language descriptions.
• Uses the Google Gemini API for text-to-code conversion.
• Provides a Flask backend with a /api/generate-mcp endpoint.
• Frontend built with React and Vite (French café-themed UI).
• Saves generated Python code and JSON configuration to files.
• Backend chooses the output directory automatically.

Use cases of MyMCP Prompt

• Quickly prototyping a custom MCP server without manual coding.
• Generating a server for a specific task by describing it in plain English.
• Learning how MCP servers are structured by reviewing generated code and config.
• Rapidly iterating on server ideas during development.

FAQ from MyMCP Prompt

What does MyMCP Prompt generate?

It generates a complete Python MCP server script and a corresponding JSON configuration file based on a user-written description.

What are the runtime dependencies?

Python packages listed in requirements.txt (Flask, Google Generative AI client, etc.) and Node.js dependencies for the frontend (Vite, React). A Google Gemini API key is also required.

How do I set the Gemini API key?

Set the environment variable GEMINI_API_KEY to your key before starting the backend. On Windows, use set GEMINI_API_KEY=your_key; on macOS/Linux, add export GEMINI_API_KEY="your_key" to your shell config.

Where are the generated files saved?

The files are saved to a directory chosen by the backend LLM (the current version does not let the user specify an output directory). The frontend displays the saved file paths.

Which LLM does MyMCP Prompt use?

Currently, only the Google Gemini API is supported. Future versions plan to integrate additional LLMs (OpenRouter, OpenAI, Anthropic, etc.).