Omrly is an AI-driven platform that transforms handwritten question drafts into structured, randomized OMR sheets, making exam preparation faster, accurate, and cheat-resistant.
Key workflow:
- Upload handwritten draft → AI uses OCR to extract text.
- AI cleans and structures questions → exports JSON.
- Secondary AI generates multiple exam variations → rearranges questions/options.
- Export print-ready PDFs or digital OMR sheets for online exams.
- 🖊️ Handwritten Question Recognition (OCR)
- 📝 AI-Powered Text Correction & Structuring
- 🔀 Question & Answer Randomization
- 🖨️ Print-Ready PDF Export
- 💻 Digital OMR Sheet for Online Exams
- 📦 Multiple Exam Versions to Reduce Cheating
| Category | Minimum Requirement | Recommended Setup |
|---|---|---|
| OS | Windows 10 / macOS / Linux | Ubuntu 22.04 LTS or macOS Monterey |
| Python | 3.9+ | 3.11+ |
| Node.js | 16+ | 18+ |
| Frontend | React + Next.js | React 18 + Next 13 + Tailwind CSS |
| Database | PostgreSQL 13 | PostgreSQL 15 with pgAdmin |
| Browser | Chrome/Firefox latest | Chrome latest |
| AI Libraries | TensorFlow 2+, OpenCV, pytesseract | TensorFlow 2.12+, OpenCV 4.8+, EasyOCR |
| PDF Library | reportlab / fpdf2 | reportlab latest, fpdf2 latest |
| Memory | 8 GB RAM | 16+ GB RAM for faster AI processing |
- Clone Repo
git clone https://github.com/YOUR_GITHUB_USERNAME/omrly-ai.git
cd omrly-ai- Backend Setup
cd backend
python -m venv venv
source venv/bin/activate # Linux/macOS
venv\Scripts\activate # Windows
pip install -r requirements.txt
uvicorn main:app --reload- Frontend Setup
cd frontend
npm install
npm run dev- Visit
http://localhost:3000to access Omrly.
- Docker: Use
docker-compose up --build(backend + frontend). - Cloud: Deploy backend on AWS/GCP/Azure; frontend on Vercel/Netlify.