Skip to content
How it works

A session is a conversation, not a form

You talk, it listens, it pushes back. Nothing is typed except your resume upload. A round runs five minutes on the free plan, fifteen on Pro, and ends with a scorecard.

01

Upload your resume

A PDF, once. It's parsed for projects, roles and claims, and that becomes the question bank for the Resume Grill round. Skip it and you still get HR and CS Fundamentals.

02

Pick a round

HR / Behavioural, Resume Grill, or CS Fundamentals. One interviewer runs all three — the questions change, the pressure doesn't.

03

Talk

Real-time voice, both ways. It waits when you pause to think instead of cutting you off mid-thought, and it follows up when an answer stays vague. If you ramble, it interrupts — the same way a tired interviewer on their fourth call of the day would.

04

Get the scorecard

Delivery numbers, a score out of 10, a rewrite of each answer, structure notes and red flags. It lands seconds after the round ends, and it stays in your history so the next one can be compared to it.

Under the hood

Two different machines produce the feedback

Counting is not judgement. Keeping them apart is what makes the scorecard trustworthy.

Deterministic · in code
  • Filler-word counts, straight off the transcript
  • Words per minute, measured against time actually spent talking
  • Total speaking time versus session length
  • Never asked of the model, so the numbers can't be hallucinated
Judgement · LLM, schema-validated
  • A score out of 10 for the round
  • Per answer: what you said, then what a strong answer sounds like
  • Structure notes — did the answer have a shape
  • Red flags worth fixing before a real screen

A nervous pause is a pause. It is not slow speech, and the scorecard doesn't confuse the two.

Ready to be interrupted?

Closed beta on Android. Invites go out in batches.

App screenshot — round select