Project 03
AIprep
Voice-based mock interview platform coordinating Vapi, Firebase, and Gemini across real-time interviews, transcript processing, structured feedback, authentication, and progress analytics.
- Next.js
- TypeScript
- Firebase
- Vapi AI
- Google Gemini
- Firestore
- Firebase Storage
Year: 2026
Overview
AIprep coordinates real-time Vapi voice interviews with a Next.js and TypeScript application. Firebase authentication and server-side session cookies protect access, while Firestore persists interviews and analytics.
Authenticated Vapi webhooks and duplicate generation prevention support the interview workflow. Captured transcripts feed Gemini structured feedback, with failure/fallback handling and Firebase Storage profile uploads.
Engineering Highlights
- Firebase authentication with server-side session cookies.
- Real-time Vapi voice interview workflow and authenticated webhook processing.
- Duplicate interview-generation prevention.
- Transcript capture and Gemini structured feedback.
- Firestore-backed user analytics.
- Firebase Storage profile uploads.
Architecture
- User
- Vapi Voice Session
- Transcript
- Next.js Server
- Firebase / Firestore
- Gemini
↓
Structured Feedback + Analytics
Reliability & Testing
- Authenticated webhook handling
- Duplicate generation prevention
- Failure and fallback handling
Screenshots
