🎓 GenCertQuiz: Transform Textbooks into Professional Exams with Multi-Agent AI

Don't just read the textbook—master it. GenCertQuiz is the open-source engine powering Hiiragi Exam. It uses a sophisticated RAG-powered Multi-Agent system to analyze your study materials and generate high-quality, exam-style quizzes that mimic real certification standards.

🌐 See it in Action

Check out the official platform built on top of this repository: Hiiragi Exam

---

🌟 Why GenCertQuiz?

Most AI quiz generators just rephrase text. GenCertQuiz acts like a professional exam board:

• The Researcher deep-dives into facts.
• The Style Analyzer mimics the tone of specific exams (e.g., AWS SAA-C03).
• The Critic ensures no hallucinations and high academic rigor.

🚀 Key Features

• 🤖 Multi-Agent RAG Pipeline

  • Researcher Agent: Extracts core facts and definitions from uploaded textbooks.
  • Style Analyzer: Analyzes uploaded exam papers to extract question patterns, difficulty, and tone.
  • Psychometrician Agent: Drafts questions that combine factual accuracy with exam-style formatting.
  • Critic Agent: Reviews and refines questions to ensure quality and remove hallucinations.

• 📚 Content Management

  • PDF Ingestion: Upload textbooks and exam papers (text & vision analysis).
  • File Management: View and delete uploaded files.
  • Vector Search: Semantic retrieval using pgvector.

• 📝 Interactive Quiz

  • Generate quizzes with adjustable difficulty.
  • Real-time feedback and explanations.

🛠️ Tech Stack

• Backend: FastAPI, Python 3.11+, Pydantic AI Agents
• AI/ML: OpenAI GPT-4o, Anthropic Claude 3.5 Sonnet
• Database: PostgreSQL with pgvector
• Frontend: Next.js 14, React, Tailwind CSS, Lucide Icons
• Infrastructure: Docker Compose

🏗️ Architecture

graph TD
    User[User] -->|Uploads PDF| InpestAPI[Ingest API]
    User -->|Requests Quiz| GenAPI[Generate API]
    
    subgraph "Data Layer"
        Postgres[(PostgreSQL + pgvector)]
    end
    
    subgraph "Agentic RAG Engine"
        Researcher[Researcher Agent]
        StyleAnalyzer[Style Analyzer]
        Psychometrician[Psychometrician Agent]
        Critic[Critic Agent]
    end
    
    InpestAPI -->|Chunks & Embeds| Postgres
    InpestAPI -->|Extracts Style| StyleAnalyzer
    StyleAnalyzer -->|Caches Profile| Postgres
    
    GenAPI --> Researcher
    Researcher -->|Retrieves Facts| Postgres
    Researcher -->|Facts| Psychometrician
    
    StyleAnalyzer -->|Style Profile| Psychometrician
    
    Psychometrician -->|Drafts Question| Critic
    Critic -->|Feedback| Psychometrician
    Critic -->|Approved Question| User

Loading

⚡ Quick Start (Docker)

The easiest way to run the application is using Docker Compose, which handles the frontend, backend, and database.

Prerequisites

• Docker & Docker Compose
• API Keys (OpenAI, Anthropic)

1. Configure Environment

Create a .env file in the backend/ directory (or rely on the default docker-compose.yml mapping):

# Copy example env in backend
cp backend/.env.example backend/.env
# Edit backend/.env to add your API keys

2. Run with Docker Compose

docker-compose up --build

Access the application:

• Frontend: http://localhost:3000
• Backend API Docs: http://localhost:8000/docs

---

🛠️ Development Setup (Manual)

If you want to run services individually for development:

1. Database

docker-compose up -d db

2. Backend

cd backend
python3 -m venv .venv
source .venv/bin/activate
pip install -e .
cp .env.example .env  # Add your API keys
uvicorn main:app --reload --port 8000

3. Frontend

cd frontend
npm install
npm run dev