🎓 ExamAI — AI-Powered Exam Evaluator

A full-stack application that uses Groq AI (LLaMA) to automatically evaluate student exam answers and provide detailed feedback, scores, and grades.

Features

• 4 answer types: Short text, long essay, MCQ, and code answers
• AI evaluation: Groq (LLaMA) grades open-ended answers with detailed feedback, strengths, and improvement suggestions
• Role-based access: Separate teacher and student flows
• Student dashboard: View available exams and submission history
• Teacher panel: Create/publish exams with model answers to guide AI grading
• Timed exams: Countdown timer with auto-submit on expiry
• Rich results page: Score ring, per-question AI breakdown, grade distribution by type

Tech Stack

Layer	Technology
Frontend	React 18 + Vite + React Router
Backend	Node.js + Express
Database	MongoDB + Mongoose
AI	Groq (LLaMA 3.3 70B)
Auth	JWT + bcryptjs

Project Structure

exam-evaluator/
├── client/          # React frontend
│   ├── src/
│   │   ├── pages/       Login, Register, StudentDashboard, ExamPage, ResultPage, TeacherDashboard, CreateExam
│   │   ├── components/  Navbar, FeedbackCard
│   │   ├── context/     AuthContext
│   │   └── api/         axios instance
│   └── index.html
│
└── server/          # Express backend
    ├── models/      User, Exam, Submission
    ├── routes/      auth, exams, submissions
    ├── middleware/  JWT auth
    ├── services/    evaluator.js (Groq AI)
    └── index.js

Setup & Run

Prerequisites

• Node.js 18+
• MongoDB (local or Atlas)
• Groq API key (free at console.groq.com)

1. Backend

cd server
npm install
cp .env.example .env

Edit .env:

PORT=5000
MONGO_URI=mongodb://localhost:27017/exam-evaluator
JWT_SECRET=your_secret_key_here
GROQ_API_KEY=gsk_your-key-here
CLIENT_URL=http://localhost:5173

npm run dev   # starts on http://localhost:5000

2. Frontend

cd client
npm install
npm run dev   # starts on http://localhost:5173

Open http://localhost:5173

API Endpoints

Auth

• POST /api/auth/register — register (name, email, password, role)
• POST /api/auth/login — login (email, password)
• GET /api/auth/me — get current user

Exams

• GET /api/exams — list all published exams (students) / own exams (teachers)
• GET /api/exams/:id — get exam by ID
• POST /api/exams — create exam (teacher only)
• PUT /api/exams/:id — update exam (teacher only)
• DELETE /api/exams/:id — delete exam (teacher only)

Submissions

• POST /api/submissions — submit exam (triggers AI evaluation)
• GET /api/submissions/my — student's own history
• GET /api/submissions/:id — get submission details
• GET /api/submissions/exam/:examId — all submissions for an exam (teacher only)

How AI Evaluation Works

1. MCQ — rule-based, instant (no API call)
2. Short text / Essay / Code — sends question, model answer, and student answer to Groq
3. Groq returns JSON: { score, feedback, strengths[], improvements[] }
4. After all questions, Groq generates an overall performance summary
5. Results are saved to MongoDB and shown to the student

Grading Scale

Grade	Percentage
A	≥ 90%
B	≥ 75%
C	≥ 60%
D	≥ 45%
F	< 45%