Examina Ai

Examina AI is a production-grade AI platform designed for educational institutions and corporate training centers. It transforms raw study materials into structured, verified examination sets with support for institutional export formats like Moodle XML.

🌟 Premium Features

🧠 Intelligent Generation Engine

• Multi-Account Token Rotation: Automatically rotates through multiple Hugging Face API tokens to bypass rate limits and credit depletion (Status 402 detection).
• Practical Application Focus: Specialized prompt engineering for Scenario-Based Problems, ensuring exams test understanding, not just memorization.
• Dynamic Question Types: Supports MCQs, True/False, Short Answers, and Essay questions with customizable weights.

🛡️ Enterprise Infrastructure

• Secure Admin Portal: Centralized hub for generating Institutional Access Keys (OTPs), tracking demo usage, and managing platform security.
• Automated Lifecycle Management: Integrated background cleanup service that purges temporary assets (uploads/outputs) every 10 minutes to maintain server hygiene.
• IP-Based Rate Limiting: Intelligent demo tracking to prevent abuse while allowing genuine potential users to trial the engine.

💎 Premium UI/UX

• Vercel-Style Aesthetic: Professional dark-mode design system featuring Glassmorphism, refined typography (Outfit & Inter), and smooth micro-animations.
• Real-Time System Health: Live monitoring of backend connectivity with automatic "Wake-up" protocols for serverless deployments.

🏗 System Architecture

graph TD
    User((User)) -->|Uploads PDF/Docx| FE[Next.js Frontend]
    FE -->|API Request| BE[FastAPI Backend]
    BE -->|Task Queue| Redis[(Redis)]
    BE -->|Store Metrics| DB[(PostgreSQL)]
    BE -->|Batch Processing| LLM[HF Router / Qwen 2.5]
    LLM -->|Rotates Tokens| BE
    BE -->|Cleanup Task| Disk[Local Storage]

🚀 Deployment Guide (Render.com)

Examina AI is optimized for deployment as two separate services on Render.

1. Backend Deployment (Web Service)

1. Create a new Web Service on Render and connect your repository.
2. Environment: Python
3. Build Command: pip install -r backend/requirements.txt
4. Start Command: gunicorn -w 4 -k uvicorn.workers.UvicornWorker backend.main:app
5. Environment Variables:
   • HF_TOKENS: token1,token2,token3 (Comma-separated)
   • ADMIN_EMAIL: Your admin email
   • ADMIN_PASSWORD: Secure admin password
   • PORT: 8000

2. Frontend Deployment (Web Service)

1. Create a new Web Service (not Static Site, for Next.js features) on Render.
2. Root Directory: frontend/
3. Environment: Node
4. Build Command: npm install && npm run build
5. Start Command: npm run start
6. Environment Variables:
   • NEXT_PUBLIC_API_URL: https://your-backend-url.onrender.com/api/v1 (The URL Render gave you for the backend)

🛠 Local Setup

Prerequisites

• Docker & Docker Compose

Installation

1. Clone & Enter:

   git clone https://github.com/RasikhAli/Examina-AI.git
   cd Examina-AI
   

2. Configure .env: Copy .env.example to .env and fill in your keys.
3. Launch:

   docker-compose up --build
   

📁 Minute Folder Details

• backend/app/core/llm_engine.py: The heart of AI generation with token rotation logic.
• backend/app/tasks/cleanup.py: Background worker for file retention policy.
• backend/app/core/payment_state.py: Global state for OTPs and demo restrictions.
• frontend/src/app/globals.css: The entire design system and glassmorphism tokens.
• frontend/src/app/admin/: Secure management portal.

👨‍💻 Author

Rasikh Ali
AI Engineer & Full-Stack Developer
LinkedIn | GitHub

⚖️ License

MIT License - Developed for Cognivex / Rasikh Ali.