MCPTeachingAidSystem

# AI Teaching Assistant System Based on MCP with Built-in Agents This is an intelligent teaching assistant system based on MCP (Machine Conversation Protocol), integrating large model capabilities and intelligent agent technology to support teachers in the classroom. ## Project Overview This system is built on the MCP protocol and large language models, creating an intelligent teaching assistant tool that helps teachers with knowledge queries, course planning, student tutoring, and material organization. Through conversational interaction, it provides intelligent teaching support for educators. ## Tech Stack - **Frontend**: Vue 3 + Tailwind CSS + Element Plus - **Large Model Provider**: Deepseek V3 - **Agent Implementation**: OpenAI Agent API, MCP Python API - **Backend**: Quart + CORS - **Data Storage**: MySQL - **Browser Interaction**: Puppeteer ## Features - **Intelligent Conversational Tutoring**: The AI assistant can answer various academic questions, providing clear explanations and guidance. - **Personalized Learning Paths**: The system analyzes learning situations and recommends suitable learning content and pace. - **Real-time Feedback and Assessment**: Instant feedback after exercises to understand mastery levels and areas for improvement. - **Multimedia Learning Resources**: Provides videos, charts, and interactive content to meet different learning styles. - **Document Generation and Export**: Supports PDF file generation for easy saving and sharing of learning content. - **Web Information Retrieval**: Access online resources and information through browser services. ## Integrating MCP Services The system integrates the following MCP servers: - **SQL MCP**: Provides database query capabilities for accessing student information and learning data. - **Browser MCP**: Supports web interaction and information retrieval, allowing access to online educational resources. - **Filesystem MCP**: Provides file read and write functionalities for managing lesson plans and learning materials. - **PDF MCP**: Supports exporting conversation content and learning reports as PDF files. ## Startup Instructions ### Environment Requirements - Python 3.13+ - Node.js 18+ ### Backend Setup 1. Install Python dependencies: ```bash uv venv .venv source .venv/bin/activate # Linux/macOS .venv\Scripts\activate # Windows uv pip install -e . ``` 2. Configure environment variables: The project includes a basic `.env` file; you can modify the settings in `mcp/.env` as needed. 3. Start the MCP service: ```bash cd mcp python main.py ``` 4. Start the backend service: ```bash python run_server.py ``` ### Frontend Setup 1. Install Node.js dependencies: ```bash npm install ``` 2. Start the development server: ```bash npm run dev ``` 3. Build the production version: ```bash npm run build ``` ## Project Structure - **src/components**: Vue components, including the chat interface, login page, etc. - **/mcp**: MCP server implementation - main.py: Core server and Agent integration - pdf_server.py: PDF generation service - filesystem-server.py: File management service - browser-server.py: Web browsing service - local_web_server.py: Local web redirection service - **src/router**: Frontend routing configuration - **src/agents**: Agent implementation and control logic - **src/assets**: Static resources - **doc**: Lesson plans and teaching materials ## Testing and Debugging The system provides testing tools to verify the functionality of each module: - `test_agent_connection.py`: Tests the connection to the Agent - `test_mysql.py`: Validates the database connection ### Question Bank Generation and Import The question bank generation and import functionality can be achieved through the `Question Bank Generator` module. Users can generate different types of questions as needed and import them into the system for use.