MCP-email-fetch

# MCP Email Fetch ## Overview **MCP Email Fetch** is an AI-powered email assistant that integrates the **Model Context Protocol (MCP)** to fetch, process, rank, and summarize emails. It leverages LLM-based sentiment analysis and a vector store for efficient retrieval. The project consists of a **Python-based backend** and an **MCP server** for seamless interaction. ## Features - **Email Fetching**: Retrieves emails from Gmail. - **AI-Powered Summarization**: Extracts key insights from emails. - **Sentiment Analysis**: Determines email sentiment for prioritization. - **Email Ranking**: Uses an LLM-based ranking system. - **FAISS Vector Store**: Stores email embeddings for efficient retrieval. - **Automatic Replies**: Generates AI-powered responses to emails. - **Dashboard API**: Provides endpoints for user interaction. ## Project Structure ``` └── hiteshydv001-mcp-email-fetch/ ├── backend/ │ ├── main.py # Main entry point for backend API │ ├── requirements.txt # Dependencies for backend │ ├── ai/ │ │ ├── email_ranker.py # AI-powered email ranking │ │ ├── llm_handler.py # Handles interactions with LLM │ │ └── sentiment_analysis.py # Sentiment classification │ ├── config/ │ │ ├── config.py # General configuration │ │ ├── logging_config.py # Logging configuration │ │ ├── mcp_config.py # MCP-specific settings │ │ ├── token.json # Google API authentication token │ │ └── secrets.json # Google API client secrets │ ├── database/ │ │ ├── db.py # Database connection │ │ └── models.py # Database models │ ├── email_processing/ │ │ ├── email_reply.py # AI-generated replies │ │ ├── email_summarizer.py # Summarizes email content │ │ └── gmail_fetch.py # Fetches emails from Gmail │ ├── mcp/ │ │ ├── faiss_vectorstore.py # FAISS-based email storage │ │ └── mcp_retriever.py # Retrieves relevant emails │ ├── routes/ │ │ ├── dashboard_routes.py # API routes for dashboard │ │ └── email_routes.py # API routes for emails │ └── utils/ │ ├── logger.py # Logging utilities │ ├── preprocess.py # Data preprocessing │ └── token_manager.py # Token authentication └── mcp-server/ ├── package-lock.json ├── package.json ├── tsconfig.json # TypeScript config for MCP server ├── dist/ │ └── index.js # Compiled MCP server └── src/ └── index.ts # MCP server implementation ``` ## Installation ### Backend Setup 1. **Clone the repository**: ```sh git clone https://github.com/Hiteshydv001/mcp-email-fetch.git cd mcp-email-fetch/backend ``` 2. **Create a virtual environment**: ```sh python -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate ``` 3. **Install dependencies**: ```sh pip install -r requirements.txt ``` 4. **Set up environment variables and Google API authentication**: - Obtain **Google Cloud Email API credentials**. - Download `token.json` and `secrets.json` from Google Cloud Console. - Place them inside the `config/` directory. - Ensure `gmail_fetch.py` uses these credentials. - Create a `.env` file in the `backend/` directory and add: ```ini MONGODB_URI=your_mongodb_connection_string GEMINI_API_KEY=your_gemini_api_key GMAIL_SECRET_KEY=your_gmail_secret_key DEBUG=True ``` 5. **Run the backend server**: ```sh python main.py ``` ### MCP Server Setup 1. **Navigate to the `mcp-server` directory**: ```sh cd ../mcp-server ``` 2. **Install dependencies**: ```sh npm install ``` 3. **Run the MCP server**: ```sh npm run start ``` ## Usage - Use the **API endpoints** to fetch, summarize, and reply to emails. - Interact via a **dashboard** to view email sentiment and ranking. - Automate email responses with **LLM-generated replies**. ## Contributing Feel free to contribute by submitting issues or pull requests. ## License This project is licensed under the MIT License. --- 🚀 **Developed by Hitesh Kumar**