Mockr

Overview

AI interview practice with structured feedback. JD-targeted or general mode, voice or typed answers, four-axis scoring rubric. Question prefetching keeps the next question instant - no waiting after feedback.

What I shipped

• JD-targeted mode parses a job description and generates role-specific questions; general mode for broad practice.
• Voice or typed answers with a four-axis rubric: structure, specificity, signal, and delivery.
• Question prefetching: the next question is generated while the user reviews feedback, so there's no spinner mid-session.
• Browser-only PDF parsing via unpdf. No JD content ever leaves the client until inference.

Stack

• Next.js 16
• React 19
• TypeScript
• LLM API
• unpdf