PrepAI — AI-Powered Interview Preparation

PrepAI analyzes your resume and a target job description using AI to generate a personalized interview plan: technical questions, behavioral questions, skill gap analysis, a day-by-day preparation roadmap, and a tailored resume PDF.

---

Tech Stack

Backend — Node.js, Express.js, MongoDB (Mongoose), JWT, Multer, pdf-parse, Puppeteer, Groq SDK

Frontend — React 19, Vite, React Router v7, Axios, SCSS

AI — Groq API (Llama 3.3 70B)

---

Project Structure

YT-GENAI/
├── Backend/
│   ├── src/
│   │   ├── config/         # MongoDB connection
│   │   ├── controllers/    # Route handlers
│   │   ├── middlewares/    # Auth (JWT) + File upload (Multer)
│   │   ├── models/         # Mongoose schemas
│   │   ├── routes/         # Express routers
│   │   └── services/       # AI (Groq) + PDF generation (Puppeteer)
│   ├── server.js
│   └── package.json
├── Frontend/
│   ├── src/
│   │   ├── features/
│   │   │   ├── auth/       # Login, Register, Protected route
│   │   │   └── interview/  # Home, Interview report, hooks, API
│   │   ├── App.jsx
│   │   └── app.routes.jsx
│   └── package.json
├── .gitignore
├── netlify.toml
├── render.yaml
└── README.md

---

Local Setup

1. Clone the repo

git clone https://github.com/your-username/prepai.git
cd prepai

2. Backend

cd Backend
npm install

Create Backend/.env:

MONGO_URI=your_mongodb_connection_string
JWT_SECRET=any_random_secret_string
GROQ_API_KEY=your_groq_api_key
PORT=3000

npm run dev

3. Frontend

cd Frontend
npm install
npm run dev

Frontend: http://localhost:5173 — Backend: http://localhost:3000

---

Environment Variables

Variable	Description
MONGO_URI	MongoDB Atlas connection string
JWT_SECRET	Random secret for signing JWTs
GROQ_API_KEY	Free API key from console.groq.com

---

Deployment

Step 1 — Push to GitHub

Run these commands from inside your YT-GENAI project folder:

git init
git add .
git commit -m "initial commit"
git branch -M main
git remote add origin https://github.com/your-username/prepai.git
git push -u origin main

Step 2 — Deploy Backend on Render

1. Go to render.com → New → Web Service
2. Connect your GitHub repo
3. Configure:
   • Root Directory: Backend
   • Build Command: npm install
   • Start Command: node server.js
4. Add environment variables: MONGO_URI, JWT_SECRET, GROQ_API_KEY
5. Deploy — copy your backend URL (e.g. https://prepai-backend.onrender.com)

Step 3 — Update Frontend API URLs

Before deploying frontend, update the baseURL in both files:

Frontend/src/features/auth/services/auth.api.js:

baseURL: "https://prepai-backend.onrender.com"

Frontend/src/features/interview/services/interview.api.js:

baseURL: "https://prepai-backend.onrender.com"

Commit and push:

git add .
git commit -m "update api base url for production"
git push

Step 4 — Deploy Frontend on Netlify

1. Go to netlify.com → Add new site → Import from Git
2. Connect your GitHub repo
3. Configure:
   • Base directory: Frontend
   • Build command: npm run build
   • Publish directory: Frontend/dist
4. Deploy

The netlify.toml file handles React Router redirects automatically — no extra setup needed.

---

API Reference

Auth

Method	Endpoint	Description
POST	/api/auth/register	Register { username, email, password }
POST	/api/auth/login	Login { email, password }
GET	/api/auth/logout	Logout
GET	/api/auth/get-me	Get current logged-in user

Interview

Method	Endpoint	Description
POST	/api/interview/	Generate report (multipart: resume PDF + jobDescription + selfDescription)
GET	/api/interview/	Get all reports for current user
GET	/api/interview/report/:id	Get single report by ID
POST	/api/interview/resume/pdf/:id	Download tailored resume as PDF

---

Features

• Upload PDF resume or type a self-description
• Paste any job description
• AI generates:
  • Match score (0–100%)
  • 5 technical questions with model answers
  • 5 behavioral questions with model answers
  • Skill gaps with severity (low / medium / high)
  • 7-day preparation plan
• Download a tailored resume PDF built for the target role
• All past reports saved and viewable from home page

---

Known Limitations

• Render free tier sleeps after 15 min idle — first request takes ~30s to wake up
• Puppeteer PDF generation is CPU-heavy; may be slow on Render free tier
• Groq free tier: 14,400 requests/day, 30 req/min on Llama 3.3 70B

---

Author

Harsh Yadav GitHub: github.com/harsh-0905 Portfolio: portfolio-7ivq.vercel.app LinkedIn: linkedin.com/in/harshyadav95-dev