Chiron-AI (TATopia)

The best personalized AI teaching assistant in the world.

Chiron-AI is an intelligent educational technology platform that provides adaptive, personalized tutoring support for students while empowering professors with tools to manage courses and learning materials. Built with a multi-agent AI architecture based on Bloom's taxonomy, it delivers context-aware assistance tailored to each student's learning needs.

Visit us at: chiron-ta.com

---

Table of Contents

• Features
• Architecture
• Tech Stack
• Getting Started
• Project Structure
• API Documentation
• Deployment
• Contributing

---

Features

For Students

• Intelligent AI Tutoring: Chat with an AI teaching assistant powered by specialized expert agents
• PDF Document Integration: View and interact with course materials while studying
• Conversation Management: Access chat history and resume previous learning sessions
• Personalized Learning: AI adapts to your learning style and needs
• Real-time Responses: Get immediate help with academic questions

For Professors

• Course Management: Create and manage classes with student enrollment
• Material Upload: Upload and organize PDF course materials
• AI-Powered Summarization: Automatically generate summaries of course materials
• Student Analytics: Track student engagement and learning characteristics
• Model Configuration: Customize AI assistant behavior for your courses

For Administrators

• System Administration: Full user and role management capabilities
• Entity Management: Create and assign users, classes, and materials
• Platform Oversight: Monitor and manage the entire educational ecosystem

---

Architecture

Multi-Agent AI System

Chiron-AI uses a sophisticated multi-agent architecture based on Bloom's taxonomy of learning:

[Supervisor Agent] → Routes questions to appropriate specialist
         ↓
    ┌─────────────────────────────────────────┐
    │  7 Expert Agents (Bloom's Taxonomy):    │
    ├─────────────────────────────────────────┤
    │ 1. Recall Agent (basic facts)           │
    │ 2. Conceptual Clarification Agent       │
    │ 3. Application/Procedural Agent         │
    │ 4. Analysis & Reasoning Agent           │
    │ 5. Evaluation & Critique Agent          │
    │ 6. Creation & Synthesis Agent           │
    │ 7. Study Skills Agent                   │
    └─────────────────────────────────────────┘
         ↓
    [Semantic Search Tools + Context Retrieval]
         ↓
    [Evaluation Agent] → Quality control
         ↓
    [Response to Student]

Each agent specializes in a different cognitive level, ensuring students receive appropriate support whether they're memorizing facts or creating new synthesis.

---

Tech Stack

Backend

• Framework: FastAPI (Python)
• AI Orchestration: LangGraph for multi-agent workflows
• LLM Models: OpenAI (GPT-4o, GPT-4o-mini, o1-mini)
• Databases:
  • MongoDB (user data, classes, materials, chats)
  • Qdrant (vector database for semantic search)
• Embeddings: VoyageAI (voyage-3-large)
• Object Storage: Wasabi S3 (S3-compatible)
• Authentication: JWT tokens (HS256)
• Async Runtime: Motor (async MongoDB driver)

Frontend

• Framework: Next.js 15 (React 18)
• Language: TypeScript
• UI Components: Material-UI (MUI), Tailwind CSS
• PDF Handling: React PDF Viewer, PDF.js
• Math Rendering: KaTeX
• Markdown: React Markdown
• Data Fetching: TanStack Query (React Query)
• Authentication: JWT with Google OAuth

Infrastructure

• Backend Hosting: DigitalOcean
• Frontend Hosting: Vercel
• Reverse Proxy: NGINX
• Containers: Docker
• Debugging: LangSmith
• Version Control: Git with Git LFS for large assets

---

Getting Started

Prerequisites

• Python 3.8+
• Node.js 18+
• MongoDB instance (local or Atlas)
• Qdrant cloud account
• OpenAI API key
• VoyageAI API key
• Wasabi S3 credentials

Backend Setup

1. Navigate to the backend directory:

cd backend

2. Create a virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Create .env file from example:

cp .env.example .env

5. Configure environment variables in .env:

OPENAI_API_KEY=your_openai_key
VOYAGEAI_API_KEY=your_voyage_key
MONGODB_URI=your_mongodb_uri
QDRANT_URL=your_qdrant_url
QDRANT_API_KEY=your_qdrant_key
JWT_SECRET_KEY=your_secret_key
WASABI_ACCESS_KEY=your_wasabi_key
WASABI_SECRET_KEY=your_wasabi_secret

6. Run the development server:

uvicorn app.main:app --reload

The API will be available at http://localhost:8000

Frontend Setup

1. Navigate to the frontend directory:

cd tatopia-web

2. Install dependencies:

npm install

3. Create .env.local file:

NEXT_PUBLIC_API_URL=http://localhost:8000
NEXT_PUBLIC_GOOGLE_CLIENT_ID=your_google_client_id

4. Run the development server:

npm run dev

The application will be available at http://localhost:3000

---

Project Structure

Chiron-AI/
├── backend/                     # Python FastAPI application
│   ├── app/
│   │   ├── main.py             # Application entry point
│   │   ├── core/               # Core configuration
│   │   │   ├── settings.py     # Model & API settings
│   │   │   ├── config.py       # Application config
│   │   │   └── database.py     # Database connections
│   │   ├── routers/            # API endpoints
│   │   │   ├── chatbot/        # Chat API routes
│   │   │   ├── common/         # Auth & common routes
│   │   │   └── db/             # Database routes
│   │   ├── services/           # Business logic
│   │   │   ├── chat/           # AI chat services
│   │   │   │   ├── langgraph/  # LangGraph state machine
│   │   │   │   ├── supervisor_agent/
│   │   │   │   ├── expert_agent/
│   │   │   │   └── evaluation_agent/
│   │   │   └── db/             # Database services
│   │   ├── models/             # Pydantic models
│   │   ├── middlewares/        # Auth middleware
│   │   └── utils/              # Utilities
│   └── requirements.txt
│
└── tatopia-web/                # Next.js frontend
    ├── src/
    │   ├── app/                # Next.js pages
    │   │   ├── login/
    │   │   ├── student/
    │   │   ├── professor/
    │   │   └── admin/
    │   ├── components/
    │   │   └── ui/
    │   │       ├── student/
    │   │       ├── professor/
    │   │       ├── admin/
    │   │       └── common/
    │   └── middleware.ts       # Auth middleware
    ├── package.json
    └── next.config.ts

---

API Documentation

Key Endpoints

Chat Operations

• POST /chat - Send message and get AI response
• POST /generate_thread_id - Create new conversation thread
• POST /prompt_polishing - Refine user prompts

Database Operations

• POST/GET /students/* - Student management
• POST/GET /professors/* - Professor management
• POST/GET /classes/* - Class management
• POST/GET /materials/* - Course materials
• POST/GET /chats/* - Chat history

File Operations

• POST /upload - Upload PDF materials (max 100MB)
• POST /add_pdf_to_storage - Process PDF into vector embeddings
• GET /download - Retrieve stored materials

Storage

• POST /qdrant/* - Vector database operations
• POST /s3/* - Object storage operations

---

Deployment

Production Architecture

• Backend: DigitalOcean Droplet with Uvicorn + NGINX reverse proxy
• Frontend: Vercel (automatic deployments from git)
• Database: MongoDB Atlas
• Vector DB: Qdrant Cloud
• Object Storage: Wasabi S3

Environment Configuration

Ensure all production environment variables are properly configured:

• API keys (OpenAI, VoyageAI)
• Database URIs (MongoDB, Qdrant)
• JWT secret keys
• Storage credentials (Wasabi)
• Domain configurations

---

Contributing

Development Workflow

1. Create a feature branch from main
2. Make your changes
3. Ensure code quality:
   • Backend: Follow PEP 8 style guide
   • Frontend: Run npm run lint
4. Test your changes thoroughly
5. Submit a pull request

Code Review Guidelines

• Write clear, descriptive commit messages
• Include tests for new features
• Update documentation as needed
• Follow existing code patterns and conventions

---

License

Proprietary - All rights reserved

---

Contact

For questions or support, please contact:

• Repository Owner: Chang Min Bark, Hung Ngo, Duy Le
• Website: chiron-ta.com

---

Built with advanced AI technology to revolutionize personalized education.