source-rashi/Evalio

Evalio

Smart Exam Evaluation System powered by AI Vision and NLP.

Overview

Evalio automates grading of subjective answers using OCR and NLP. This repository contains a minimal scaffold for a full-stack app with a Node/Express backend and a React frontend.

Quick setup (development)

1. Install dependencies for backend:

   npm install

2. Create a .env file with the following variables:

   MONGO_URI= OPENAI_KEY= CLOUDINARY_URL= GOOGLE_APPLICATION_CREDENTIALS= JWT_SECRET=

3. Start the backend (from repo root):

   npm run dev

4. Frontend is in frontend/ — see its README for steps.

What I scaffolded

• Basic Express server with Mongoose models for Teacher, Student, Exam, Question, Submission, Evaluation.
• Authentication stubs (JWT).
• API endpoints described in the project spec.
• AI grading service with support for OpenAI or Gemini, falling back to a heuristic scorer.
• Minimal React app (create-react-app style) under frontend/.

Next steps

• Configure AI provider in .env (AI_PROVIDER=openai|gemini|none) and set OPENAI_KEY or GEMINI_API_KEY.
• Add OCR integration (Cloudinary + Google Vision) to populate extractedText.
• Add tests, CI, and deployment workflows.