Practice technical interviews out loud. An AI interviewer asks role-specific questions in a natural voice, listens to your spoken answers, scores each one, and writes a final report with strengths, gaps and topics to study.
Voice in, voice out. Questions are spoken with Groq’s Orpheus voice. Answers are recorded in the browser and transcribed by Groq Whisper, so it works in any modern browser and not only Chrome.
Never goes silent. If Groq text-to-speech is unavailable (free-tier quota, rate limit, outage), the app switches to the browser’s built-in voice automatically.
Adaptive questions.openai/gpt-oss-120b on Groq writes each next question based on the role, the difficulty and what was already asked, with strict JSON outputs.
Per-answer scoring. Every answer gets overall, technical-accuracy and communication scores, plus strengths, weaknesses and feedback.
Final report and history. Each interview ends with a holistic assessment and study topics. Completed interviews are saved to your dashboard.
Accounts. Email and password sign-in, with optional Google sign-in. JWTs are kept in httpOnly cookies.
Interview integrity. Fullscreen mode plus tab-switch and focus-loss detection, with automatic disqualification after 3 violations.
Resilient sessions. Live interviews are stored in Postgres, so a server restart or redeploy doesn’t lose progress.
Screenshots
Setup
Report
Dashboard
Tech stack
Layer
Technology
Client
React 19, Vite 8, Tailwind CSS 4, Framer Motion
Server
Node.js 22, Express 5, pg, groq-sdk, multer
Database
PostgreSQL (Supabase in production; any Postgres 13+ works)
