A highly polished, resume-worthy, and production-ready full-stack AI application designed to revolutionize the way examinations are generated, taken, and evaluated. AegisExam leverages the power of Large Language Models (LLMs) to automatically generate comprehensive exams, evaluate subjective/coding answers, and provide actionable analytics for both students and teachers.
- Role-Based Access Control: Secure JWT authentication for Students, Teachers, and Admins.
- AI Exam Generation: Dynamic prompt engineering pipeline to generate MCQs, True/False, Fill in the Blanks, Theory, and Coding questions using Gemini or OpenAI APIs.
- RAG Document Indexer: Upload PDF/DOCX/TXT study materials to generate contextual, accurate questions based directly on the uploaded corpus.
- Interactive Exam Environment: Fullscreen mode, anti-cheat warnings, timed sessions, code editors for programming questions, and mark-for-review capabilities.
- AI Answer Evaluation: Semantic grading for theory questions and sandbox execution simulation for coding challenges, returning intelligent feedback and missing keywords.
- Advanced Analytics: Comprehensive Recharts-powered dashboard showing historical performance, score trends, and AI-recommended study topics.
- PDF Reports: High-resolution, professional PDF exports of student evaluation reports using HTML2Canvas and jsPDF.
- Modern UI/UX: Built with TailwindCSS, Framer Motion, and Shadcn UI with responsive design, glassmorphism aesthetics, and dark mode support.
- Framework: Next.js 13 (App Router) & React 18
- Styling: Tailwind CSS, Framer Motion, Shadcn UI
- State Management: Zustand
- Data Fetching: Axios, React Query
- Charts: Recharts
- Framework: Python FastAPI
- Database: PostgreSQL (with SQLite zero-config fallback)
- ORM: SQLAlchemy & Alembic
- AI Integration: Google GenAI SDK (Gemini 2.5 Flash), OpenAI
- Security: Passlib (Bcrypt), PyJWT, FastAPI CORS Middleware
- Node.js (v18+)
- Python (3.10+)
- (Optional) Docker
cd backend
python -m venv venv
# Activate virtual environment (Windows: .\venv\Scripts\activate | Mac/Linux: source venv/bin/activate)
pip install -r requirements.txt
# Set up your environment variables
echo "GEMINI_API_KEY=your_api_key_here" > .env
# Start the server
uvicorn app.main:app --host 0.0.0.0 --port 8000cd frontend
npm install
npm run devThe platform will be available at http://localhost:3000.
For production deployment, you can use the provided Docker Compose configuration:
docker compose up -d --buildThis will automatically provision the PostgreSQL database, FastAPI backend, and Next.js frontend in isolated containers.
- Request Phase: Teacher submits exam parameters (Topic, Difficulty, Bloom Level) via the React dashboard.
- Prompt Builder: FastAPI constructs a highly-constrained system prompt dictating the exact JSON schema required.
- LLM Generation: The payload is sent to Gemini (or OpenAI). The response is parsed and validated using Pydantic schemas.
- Data Persistence: The generated exam and its questions are saved to the relational database (PostgreSQL).
- Evaluation Phase: When a student submits their attempt, subjective answers are bundled with the model answer and sent back to the LLM for semantic evaluation (assigning a float score and feedback).
Please read CONTRIBUTING.md for details on our code of conduct, and the process for submitting pull requests to us. Ensure your code is thoroughly tested (Pytest/Jest) before opening a PR.
Built as a showcase for Advanced Agentic AI Engineering.
