🎯 Interview AI — AI-Powered Interview Preparation Platform

[

An end-to-end full-stack web application that leverages Google Gemini Generative AI to analyze your resume and target job descriptions, generating tailored interview strategies, technical/behavioral question banks, skill gap analyses, day-wise roadmaps, and ATS-optimized tailored resumes — and then takes you from preparation to practice with an AI-driven mock interview module (timed & voice modes), instant answer evaluation with smart follow-ups, and a progress dashboard that tracks your readiness over time.

🔗 Live Application URL: [https://interview-ai-phi-one.vercel.app/)

---

✨ Key Features

• 📊 Smart Profile Match Scoring

  • Instant candidate-to-job match score calculation (0–100%).
  • Highlights compatibility and readiness for the target role.

• 🎯 Targeted Technical & Behavioral Question Banks

  • Curated technical questions based on the candidate's exact tech stack and JD requirements.
  • Deep behavioral questions accompanied by interviewer intent and model answers.

• 🔍 Skill Gap Analysis

  • Identifies missing or weak skills with severity indicators (Low, Medium, High).

• 🗓️ Day-Wise Preparation Road Map

  • Structured, actionable preparation schedule broken down by daily focus and tasks.

• 📄 ATS-Friendly Tailored Resume PDF Generation

  • Uses AI and Puppeteer to generate and download a clean, ATS-optimized resume tailored specifically to the job description.

• 🎤 AI-Driven Mock Interview (Timed & Voice Modes)

  • Practice real questions from your report in a full-screen mock interview session.
  • Timed mode: answer each question against a countdown timer with auto-submit.
  • Voice mode: record your spoken answer with the mic; Gemini audio input transcribes it verbatim for evaluation.

• 📝 Instant Answer Evaluation & Smart Follow-ups

  • Every practice answer is scored (0–100) against the model answer.
  • Get concise feedback, key-point coverage, 2–3 improvement tips, and smart follow-up questions that probe exactly what you missed.

• 📊 Practice Progress & Readiness Dashboard

  • Aggregates every practice/voice/timed attempt into per-report readiness scores.
  • Tracks average scores by section (technical/behavioral), a 14-day score history, practice streak, and improvement on repeated questions.

• 🔒 Secure Authentication & Route Protection

  • JWT-based authentication stored in secure HttpOnly cookies.
  • Protected frontend route guards and custom auth hooks.

---

🛠️ Tech Stack

Frontend

• Framework: React 19 + Vite
• Routing: React Router v7
• Styling: SCSS (Modern Dart Sass @use "sass:color")
• State Management: React Context API (AuthProvider, InterviewProvider)
• HTTP Client: Axios (with credentials)
• UI Kit: Reusable components (Button, Card, Tabs, Toast, Dropzone, Accordion, …) with lucide-react icons
• Media Recording: Web MediaRecorder API (useMediaRecorder hook) for voice answers

Backend

• Runtime: Node.js + Express 5
• Database: MongoDB + Mongoose ODM
• AI Integration: @google/genai (Google Gemini API — text generation & audio transcription)
• PDF Processing: pdf-parse v2 (Resume extraction) & puppeteer (Resume PDF generation)
• File Upload: Multer (Memory storage — files & audio)
• Authentication: JSON Web Tokens (jsonwebtoken) & bcryptjs

---

📁 Project Architecture

GenAi/
├── Backend/
│   ├── src/
│   │   ├── controllers/      # Auth, interview report & practice controllers
│   │   ├── middlewares/      # Auth guard & Multer (file + audio) upload middleware
│   │   ├── models/           # Mongoose schemas (User, InterviewReport, PracticeAttempt)
│   │   ├── routes/           # Express API route endpoints
│   │   └── services/         # Gemini AI (report, evaluation, transcription) & Puppeteer PDF services
│   ├── .env.example          # Template environment variables
│   ├── .gitignore
│   ├── package.json
│   └── server.js             # Server entry point
│
├── Frontend/
│   ├── src/
│   │   ├── components/       # Navbar & reusable UI kit (Button, Card, Tabs, Toast, …)
│   │   ├── features/
│   │   │   ├── auth/         # Login, Register, AuthContext, Protected routes
│   │   │   └── interview/    # Home (plan generator), Interview (report view), MockInterview, ProgressPanel
│   │   ├── styles/           # Global SCSS styles & variables
│   │   ├── app.routes.jsx    # Client-side router configuration
│   │   └── main.jsx          # React app entry point
│   ├── .env.example          # Template environment variables (VITE_API_URL)
│   ├── .gitignore
│   ├── package.json
│   └── vite.config.js
│
├── .gitignore                # Root gitignore
└── README.md

---

🚀 Getting Started

Prerequisites

• Node.js (v18 or higher)
• MongoDB (Local instance or MongoDB Atlas cluster)
• Google AI Studio API Key (Get your API Key)

---

1. Clone the Repository

git clone https://github.com/Varunmalewar/Interview-AI.git
cd Interview-AI

---

2. Backend Setup

1. Navigate to the backend directory:

   cd Backend

2. Install dependencies:

   npm install

3. Create your .env file based on .env.example:

   cp .env.example .env

4. Configure your environment variables in .env:

   MONGO_URL=mongodb+srv://<username>:<password>@cluster.mongodb.net/interview-master
   JWT_SECRET=your_super_secret_jwt_key
   GOOGLE_AI_API_KEY=your_gemini_api_key
   CLIENT_URL=http://localhost:5173   # Frontend origin for CORS (defaults to this)
   # COOKIE_SAMESITE=none             # Uncomment for cross-site HTTPS deployments (e.g. Vercel frontend + separate API host)
   # COOKIE_SECURE=true               # Defaults to true when NODE_ENV=production

5. Start the backend development server:

   npm run dev

   The backend will run on http://localhost:3000.

---

3. Frontend Setup

1. Open a new terminal and navigate to the frontend directory:

   cd Frontend

2. Install dependencies:

   npm install

3. (Optional) Create Frontend/.env to point at a deployed API in production:

   VITE_API_URL=https://your-api-host

   In development this is not needed — the Vite dev server proxies /api/* to http://localhost:3000 automatically (see vite.config.js).

4. Start the frontend development server:

   npm run dev

   The frontend will run on http://localhost:5173.

---

🌐 API Endpoints

Authentication (/api/auth)

Method	Endpoint	Description	Access
POST	/api/auth/register	Register a new user account	Public
POST	/api/auth/login	Log in and receive auth cookie	Public
GET	/api/auth/logout	Clear auth cookie	Private
GET	/api/auth/get-me	Get currently logged-in user profile	Private

Interview & AI (/api/interview)

Method	Endpoint	Description	Access
POST	/api/interview/	Generate interview report (Upload resume PDF / JD)	Private
GET	/api/interview/	Get all generated reports for the current user	Private
GET	/api/interview/report/:interviewId	Get detailed interview report by ID	Private
POST	/api/interview/resume/pdf/:interviewReportId	Generate & download tailored ATS Resume PDF	Private
POST	/api/interview/evaluate	Evaluate an answer against the model answer (score, feedback, follow-ups)	Private
POST	/api/interview/transcribe	Transcribe a spoken answer (WAV upload) via Gemini audio input	Private
GET	/api/interview/practice/stats	Aggregated practice stats & readiness for the Progress dashboard	Private
GET	/api/interview/practice/attempts	Recent practice attempts (filter by report/section/mode)	Private

---

🔒 Security Best Practices

• No Secrets in Source Control: All sensitive keys and database URLs are managed through .env and strictly ignored via .gitignore.
• HttpOnly Cookies: Prevents client-side XSS access to auth tokens.
• Input Sanitization & Schema Validation: Strict schema formatting for GenAI outputs.

---

📄 License

This project is licensed under the ISC License.