How it works
Six steps, about ten minutes
Nothing here is hidden from you. You can read the exact criteria you are scored against before you upload anything.
- 01Send something to readUpload a CV as PDF or Word. No CV, or one that undersells you? Copy the prompt we give you into ChatGPT, Claude or Gemini and paste back what it writes about you.An AI written profile is scored more sceptically than a CV, and we say so to the judge.
- 02Pick who judges youEach hiring manager reads the same CV completely differently. Open 'See how they judge' on any card to read the exact criteria before you choose.The interview questions stay hidden, otherwise the interview is a recital.
- 03Get scored and placedYou get a score out of 100, what they rated, where you lost points, and a verdict written in their voice. Then your position against everyone else that judge has scored.Under 8 candidates on a judge we show the score and hold the rank, because a position among four people means nothing.
- 04Confirm your email, onceYour full result lands in your inbox straight away. Clicking the link in it confirms the address and opens the interview.Confirm once and every judge you try after that opens immediately.
- 05Interview to move your scoreA short conversation, in that manager's voice, about the exact gaps they found. Concrete examples move your score up. Vague answers move it down.Your score is only recalculated once you have answered every question. Leaving halfway changes nothing.
- 06Share it if you want toYou get a public card showing your score and standing, on its own link, separate from your private results page.Other candidates are never named to you, and you are never named to them.
Nobody is hiring through this pageNot today. It is a way to find out where you stand and what a specific kind of manager would push back on.
The judges are instructions, not peopleEach profile imitates how a public figure has said they hire. It is not them, not affiliated with them, and not an opinion they gave.
A score is an opinion, not a verdictOne opinionated read of one document by a language model. The gaps are usually the useful part.