Mymcp Prompt

🚀 MyMCP Prompt Tool

MyMCP Prompt is a tool for generating Model Context Protocol (MCP) servers from natural language descriptions. This Minimum Viable Product (MVP) uses the Google Gemini API to convert user descriptions into functional Python MCP servers and generate corresponding JSON configurations.

🚀 Quick Start

MyMCP Prompt is a tool that enables users to generate MCP servers from natural - language descriptions. It leverages the Google Gemini API to convert these descriptions into Python code and JSON configurations.

✨ Features

• Utilizes Google Gemini API to transform user - provided natural language descriptions into functional Python MCP servers and JSON configurations.
• Consists of a Flask backend and a React frontend, providing a user - friendly web interface.

📦 Installation

1. Clone the repository:

   git clone https://github.com/AlexJ - StL/mymcp
   cd mymcp
   

2. Backend setup (root directory): Create and activate a virtual environment (uv is recommended):

   # In the project root directory (mymcp)
   uv venv
   source .venv/Scripts/activate  # For Windows systems
   # source .venv/bin/activate    # For macOS/Linux systems
   

   Install Python dependencies:

   uv pip install -r requirements.txt
   

3. Frontend setup: Navigate to the frontend - vite directory:

   cd frontend - vite
   

   Install Node.js dependencies:

   npm install
   

   Return to the root directory:

   cd ..
   

4. Set up the Gemini API key: Note: Obtain a Google Gemini API key and set it as the environment variable GEMINI_API_KEY. Do not commit the API key to the repository. Windows (Command Prompt):

   set GEMINI_API_KEY=your_api_key
   

   Windows (PowerShell):

   $env:GEMINI_API_KEY="your_api_key"
   

   macOS/Linux:

   export GEMINI_API_KEY=your_api_key
   

💻 Usage Examples

Basic Usage

1. Start the backend server: Run the following command in the root directory:

   python app.py
   

2. Start the frontend development server: Run the following command in the frontend - vite directory:

   npm run dev
   

3. Use the application:
   • Open your browser and visit http://localhost:5173 (the default frontend development server address).
   • Enter your MCP server description.
   • Click the "Generate" button to generate the code and configuration.

📚 Documentation

Project Structure

The application consists of the following parts:

• Flask Backend (root directory):
  • app.py: Main Flask application setup (CORS, blueprint registration).
  • api.py: Contains the /api/generate - mcp endpoint, which interacts with the Google Gemini API to generate server code and configuration.
  • requirements.txt: Lists Python dependencies.
• React Frontend (/frontend - vite directory):
  • Provides a web interface (frontend - vite/src/App.js) that allows users to input server descriptions.
  • Displays the generated Python code and JSON configuration.
  • Shows the location path where the generated files are saved.

Change Log

• Version 1.0.0:
  • Initial release, supporting basic MCP server generation functionality.
  • Integrated the Google Gemini API.
  • Provided a basic user interface and backend setup.

Future Features

• Support integration with additional Large Language Models (LLMs), including OpenRouter, LiteLLM, OpenAI, Anthropic, SombaNova, Cerebras, LM Studio, Ollama, and Groq.
• Support for generating tool/function calls.
• Support for generating agent prompts.
• Improve error handling and user feedback.
• Provide more complex MCP server code generation (e.g., using templates, better structure).
• Enhance the user - experience interface.
• Add unit tests for the backend and frontend.