A rep is a timed conversation, out loud, with someone who can lose interest, get distracted and say no. You are scored on how you talked — never on whether it worked. A clean rep that ends in rejection can score 92.
One rep a day free — three on your first day. No card.
She is looking for a birthday present for her sister and getting nowhere.
Six dimensions, every rep, each one quoting your own words back at you. Whether she gave you a number, agreed to anything, or walked away contributes exactly zero. A sloppy rep that got lucky scores 54.
Warmth finished at 58 and she was never armed. None of that is in the number.
Did you get a conversation started at all. An awkward opener that lands beats a polished one that never comes.
Did you ask about her, and go past the first answer. Depth, not count.
Did you use what she gave you, or wait for your turn to talk.
Did you read how interested she was, and adjust. Missing a clear no scores low.
Did you stay steady. Recovering from an awkward moment counts for more than never having one.
How it ended. Leaving warmly without pushing scores high, including when you were turned down.
Sixty per cent of the composite is measured directly from the transcript — how much you talked, how long you let a silence sit, how many of your questions were real ones. The other forty is judged against a fixed rubric that has to quote you to justify itself.
The simulator is where you can afford to be bad at this. The point of being good at it is out there, so every rep hands you something small to do in the real world and asks you to log what actually happened.
Pick a character, read the sixty-word brief, and start. Three minutes, real time, real interruptions. She has her own reason for being there and it is not you.
The composite, the six dimensions, the two you were weakest on, and one thing you actually did well — named first, quoting you, before anything critical.
A challenge graded to where you are. Ask a stranger for the time. Ask what they are reading. Compliment a choice and then leave. Nothing that depends on a yes.
Rate how nervous you expected to be, then how nervous you were. The gap between those two lines is the chart the whole product is building toward.
Every challenge is written by hand and reviewed before it ships, against one test: the worst realistic outcome is a polite no. Nothing asks you to persist after a refusal, film anybody, or make somebody who is working the subject of an exercise.
Each one holds a rung. They are not difficulty settings with the same script — they want different things, they run out of patience differently, and the top rung takes the warmth number off your screen so you have to read a person instead of a meter.
Dry, quiet, half-distracted, and in a good mood that has nothing to do with you. She is easy to talk to and very hard to lose.
Warm, dry, and slightly guarded. She will keep this going while it is worth having and let it end when it is not.
Unfailingly polite and almost impossible to read. Whether she is interested is a question she will not answer either way.
A tier opens when you score 70 or better in two reps at the tier below it. Not when you win two — winning is not a thing you can grind, and it is not the thing being measured.
Text mode runs the same characters with no microphone and no quota. It exists because a limit that breaks your streak is a limit that makes people quit, and some nights the answer to “can you talk out loud right now” is no.
Still no luck with the present, if you were wondering.
Did you try the embarrassing shelf?
I did. I bought two. One is for me.
It is deliberately the lesser thing. Warmth in text is capped below the point a rep can be won, so it cannot produce the ending a voice rep can — that is enforced in code, not in a note. It keeps the habit alive on a bad night and then hands you back to the microphone.
Four of these, and all four are load-bearing. They decide what gets built as much as they decide what gets said.
We never write your lines. Every competitor does the frightening part on your behalf, which builds nothing and collapses the moment you are in a real room.
Characters are training equipment. They are not available for open-ended chat, they do not miss you, and every session has a hard time limit.
No clinical claims, here or anywhere in the product. If you are working with a clinician on social anxiety, this is practice alongside that, never instead of it.
Sessions are bounded at PG-13. Characters are written to decline rather than play along, steering a rep there breaches the terms, and the bound is not one we are able to waive.
Read the whole position on safety and scope — including what happens if a session goes somewhere that is no longer training.
Field work is unlimited on every plan, including the free one, forever. What you pay for is voice minutes against a live character — the only part of this that costs anything to run.
For a voice rep, yes — that is the entire exercise, and it is the part no amount of reading fixes. There is also a text mode running the same characters with no microphone and no quota, for the evenings when speaking is not on. It cannot produce the ending a voice rep can, on purpose.
Session audio is stored in a private bucket that only your account can read, and it is deleted automatically thirty days after the rep. Transcripts and scores stay, because that is what progress is made of. Ask us and the audio goes sooner.
One line each, and only about the encounter — the book she could not find, where she was going. Never about how you did and never about wanting to see you again. You can clear every line in one tap.
The field challenge is still there and still free, on every plan, forever. Running out of voice minutes never breaks a streak, because a paywall that is also a churn event is a badly designed paywall.
Dating is the sharpest wedge and it is where the roster starts, but the skill underneath — starting a conversation, reading whether it is welcome, leaving well — is the same one interviews and hard conversations at work need. Those tracks run on the same engine and are not open yet.
No. Nothing in the challenge library asks you to persist past a refusal, and the scorecard rewards reading a no correctly and leaving warmly. A rep where you push scores worse, not better.
You will probably be bad at it. That is the entire reason the simulator exists, and it is the last time being bad at it costs you nothing.
Start training free