Code Rabbit - Interview AI

Code Rabbit - Interview AI is a full-stack interview-preparation app that turns a target job description and a candidate profile into a personalized interview report. Users can upload a PDF resume or provide a short self-description, then receive an AI-generated match score, question bank, skill-gap analysis, and day-by-day preparation plan.

• Account registration, login, logout, and protected routes
• Cookie-based JWT authentication with token blacklisting on logout
• PDF resume upload and text extraction
• AI-generated interview reports tailored to a job description and candidate profile
• Technical and behavioral practice questions, including answer guidance
• Match score, skill-gap severity, and a daily preparation plan
• Saved report history for each user

• Frontend: React 19, Vite, React Router, Axios, Sass
• Backend: Node.js, Express, Mongoose, Multer, JWT, bcrypt
• Database: MongoDB
• AI: Google Gemini via @google/genai

.
+-- Backend/                 # Express API, database models, and AI integration
|   +-- src/
|   |   +-- controllers/
|   |   +-- middlewares/
|   |   +-- models/
|   |   +-- routes/
|   |   `-- services/
|   `-- server.js
`-- Frontend/                # React + Vite client
    `-- src/
        +-- features/auth/
        `-- features/interview/

• Node.js 18 or later
• A MongoDB database (local or Atlas)
• A Google AI Studio API key with access to a Gemini model

git clone <repository-url>
cd interview-ai-yt

Create Backend/.env:

MONGO_URI=mongodb+srv://<username>:<password>@<cluster>/<database>
JWT_SECRET=replace-with-a-long-random-secret
GOOGLE_GENAI_API_KEY=your-google-ai-studio-api-key

# Optional. Defaults to gemini-3.5-flash.
GEMINI_MODEL=gemini-3.5-flash

Install dependencies and start the API:

cd Backend
npm install
npm run dev

The backend runs at http://localhost:3000.

In a second terminal:

cd Frontend
npm install
npm run dev

Open http://localhost:5173 in your browser.

1. Create an account or sign in.
2. Open the interview workspace.
3. Paste the job description you want to target.
4. Upload a PDF resume (up to 3 MB) and/or add a short self-description.
5. Generate the interview strategy and review the saved report.

The report-generation request accepts jobDescription, selfDescription, and a resume file. The current server parses the uploaded resume as a PDF and keeps uploads in memory only; it does not persist the original file.

npm run dev     # Start the Express server with nodemon

npm run dev     # Start the Vite development server
npm run build   # Create a production build
npm run lint    # Run ESLint
npm run preview # Preview the production build

• The frontend is currently configured to call the API at http://localhost:3000, and the backend allows requests from http://localhost:5173.
• For production, set VITE_API_URL on the frontend to the public HTTPS backend URL, then rebuild and redeploy the frontend. Make sure this points to the API deployment, not the frontend site. Set FRONTEND_URL on the backend to the exact deployed frontend origin(s) such as https://your-site.netlify.app or https://your-site.vercel.app, then redeploy the backend. Multiple origins can be comma-separated.
• Check https://your-backend-domain/health after deploying. It must return { "status": "ok" }. If it does not, the frontend will show a network error because the API is unavailable.
• To investigate production failures, open browser DevTools → Network, retry login, and inspect the failed POST /api/auth/login request. A missing status usually means DNS, HTTPS, or CORS; a 4xx/5xx response means the API was reached. Backend logs now record the request path, origin, and error message without logging credentials.
• Keep Backend/.env private. Never commit API keys, database credentials, or JWT secrets.
• Gemini API availability and model access depend on the Google account and API key being used. You can override GEMINI_MODEL if your key uses a different supported model.

This project does not currently include a license. Add one before distributing or using it as an open-source project.