IntervAI is a full-stack AI-powered mock interview platform that enables users to practise live voice-based interviews and receive structured, automated feedback on their performance.
The platform is designed to demonstrate real interview conditions, combining real-time voice interaction with AI-driven evaluation to help candidates improve communication, problem-solving, and technical clarity.
- Conducts real-time voice interviews using AI voice agents
- Supports role, experience level, and tech-stack–specific interview sessions
- Captures and streams live transcripts during the interview session
- Dynamically generates interview questions using Google Gemini 2.0 Flash
- Questions are tailored by role, experience level, interview type (technical / behavioral / mixed), and tech stack
- Ensures clean, voice-safe output formatting for spoken interviews
-
Converts interview transcripts into structured evaluations using schema-driven AI responses
-
Provides:
- Overall interview score (0–100)
- Per-category scores (communication, technical knowledge, problem-solving, cultural fit, confidence)
- Strengths and areas for improvement
- Final qualitative assessment
- Firebase Authentication for sign-up and sign-in
- Server-side session management using HTTP-only cookies via Firebase Admin SDK
- Protected routes and user-specific interview history
- View past interviews with scores and feedback summaries
- Discover and take new AI-generated interviews
- Retake interviews to track improvement over time
Frontend
- Next.js (App Router)
- TypeScript
- Tailwind CSS
- shadcn/ui
Backend / Infrastructure
- Next.js Server Actions (server-side orchestration)
- Firebase Admin SDK
- Zod (schema validation)
AI & Voice
- Google Gemini 2.0 Flash (AI question generation & evaluation)
- VAPI.ai (real-time voice AI interviews)
- Next.js App Router used for routing, layouts, and server actions
- Server Actions handle interview orchestration and AI evaluation
- Firebase Admin manages authentication and session validation
- Zod schemas ensure consistent, typed AI responses
- Optimized server–client data flow to keep interview results and dashboards responsive
Interview Setup Dashboard: Configure role, experience level, and interview type before starting a session.
Live AI Interview Session: Real-time voice interview with streamed transcripts captured during the session.
AI-Generated Feedback: Structured post-interview evaluation with scores, strengths, and improvement areas.
📌 Screenshots are provided instead of a live demo to avoid exhausting limited AI and voice API resources.
- Session cookies are HTTP-only to reduce XSS risk
- Server-side evaluation prevents client-side manipulation
- Environment-based configuration isolates API keys and secrets
- Typed validation guards against malformed AI outputs
git clone https://github.com/your-username/intervai.git
cd intervai
npm install
npm run dev