Prepforge is your intelligent job interview preparation assistant. Powered by Next.js, Firebase, TailwindCSS, Vapi AI, and Google Gemini, it offers an immersive and modern platform where users can create, conduct, and get feedback on mock interviews with AI-powered voice agents.
Prepforge is a voice-interview training web app built to help developers and job seekers experience and learn how to integrate AI into real-world applications. With a combination of modern tools and intelligent agents, Prepforge simulates realistic job interviews and provides actionable AI feedback in real time.
Whether you're learning AI integration or preparing for your next big opportunity, Prepforge offers a hands-on, dynamic environment.
| Tech | Purpose |
|---|---|
| Next.js | Full-stack React framework for web app |
| Firebase | Authentication and real-time data storage |
| TailwindCSS | Utility-first styling for modern UI |
| Vapi AI | Voice-based AI agents for interviews |
| shadcn/ui | Accessible and beautiful UI components |
| Google Gemini | LLM for feedback generation & Q&A |
| Zod | Schema validation for forms and API routes |
- Secure Sign Up / Sign In with Firebase email-password auth.
- User session management built into the app with protected routes.
- Start new interviews with a few clicks.
- Choose role-specific prompts powered by Google Gemini and Vapi voice agents.
- Experience AI-driven voice-based interviews.
- Built with Vapi's voice agents for realistic interaction.
- After the interview, get instant AI-generated feedback and improvement suggestions.
- Transcripts and scoring powered by Gemini.
- View, manage, and organize past interviews.
- Filter by role, date, or performance score.
- Clean, responsive layout using TailwindCSS and shadcn/ui.
- Smooth navigation with accessible components.
- Fully optimized for mobile, tablet, and desktop.
- Great experience across all screen sizes.
Prepforge follows a modular, scalable architecture:
/app
├── interview/ # Routes related to interviews
├── dashboard/ # User dashboard UI
├── auth/ # Login and register flows
├── layout.tsx # Shared layout with header/footer
└── page.tsx # Landing page
/components
├── ui/ # Reusable UI components (shadcn-based)
└── interview/ # Specialized interview components
/lib
├── firebase.ts # Firebase config
├── validators/ # Zod schemas
└── vapi.ts # Vapi voice agent logic
/hooks
└── useAuth.ts # Custom hooks for auth
/utils
└── helpers.ts # Utility functions
/styles
└── globals.css # Tailwind base styles
- Type-Safe: All forms and APIs use
Zodfor runtime validation. - Reusability: All UI components are modular and easily extendable.
- API Integration: Voice and text AI responses from Gemini and Vapi are handled in clean utility modules.
git clone https://github.com/your-username/prepwise.git
cd prepwisenpm installCreate a .env.local file and add your Firebase and Vapi credentials:
NEXT_PUBLIC_VAPI_WEB_TOKEN=
NEXT_PUBLIC_VAPI_WORKFLOW_ID=
GOOGLE_GENERATIVE_AI_API_KEY=
NEXT_PUBLIC_BASE_URL=
NEXT_PUBLIC_FIREBASE_API_KEY=
NEXT_PUBLIC_FIREBASE_AUTH_DOMAIN=
NEXT_PUBLIC_FIREBASE_PROJECT_ID=
NEXT_PUBLIC_FIREBASE_STORAGE_BUCKET=
NEXT_PUBLIC_FIREBASE_MESSAGING_SENDER_ID=
NEXT_PUBLIC_FIREBASE_APP_ID=
FIREBASE_PROJECT_ID=
FIREBASE_CLIENT_EMAIL=
FIREBASE_PRIVATE_KEY=npm run dev- 🌍 Multi-language support
- 📈 Interview performance analytics
- 🧑🏽💼 AI role persona customization
- 🗂 Export transcripts to PDF
- 🔔 Real-time notifications and reminders
We’d love your input! Open an issue or submit a pull request to contribute. Prepforge is a project not just to prepare for interviews—but to learn how to integrate AI into real-world apps.
MIT © 2025 - Prepforge Team