How this works
Most resume tools start with a blank page and a model that will happily write you anything. This one starts with an interview, keeps a record of what you have actually done, and refuses to put a line in a document that isn't traceable to something you said. That is slower at the beginning and it is the entire point.
1 ·Documents optional
Old resumes, cover letters, performance reviews, brag docs. The interview reads them, so you spend your time on what they don't already say. Encrypted before they're stored, and you can delete any of them.
2 ·The interview
Five conversations, each about a different thing:
- Experience — your roles and projects, as specific stories with numbers in them
- Narrative — how you got here and what makes you different
- Skills — an honest inventory, rated 1–10 by you
- Goals — what you want next, and why
- Work style — autonomy, teams, how you like to be managed
Start with Experience. One stage is enough to build from; the rest make everything better.
Things worth knowing: you can talk instead of typing on every stage but Skills, where a misheard number would become a made-up fact. You can say “deeper” or “quick” at any point and it will change how hard it digs. If there's something specific you want to be asked about, put it on the agenda from your record and it will come up in the right conversation. Nothing is lost if you close the tab — your answer is saved before the model sees it.
3 ·Your account
Two things here matter more than the rest.
Your contact details. A resume with no name on it can't be rendered, so a build is refused outright until this is filled in. It's the one hard requirement.
Fields you're targeting, and where you can work. Leave the fields unchecked and ranking ignores the category entirely — which is fine if you're open to anything, and wrong if you're not, because you'll spend your attention on roles from an industry you'd never take.
There's more here if you want it — resume design and length, cover letter style, and standing notes applied to every document (“shorter bullets”, “lead with leadership, not tools”). All of it changes emphasis. None of it changes what is claimed.
4 ·Roles
We read employers' own job boards every hour — not an aggregator — and score what's new against your record. Including titles you'd never have thought to search for.
Three tabs, because there are three different amounts of knowing:
- Raw — a word-overlap count. A rough sort, not a judgement. It reports “23 matched”, never a score out of ten.
- Scored — a model read the posting against your record and gave it a number.
- Evaluated — the full read, against your whole record, with the reasoning written out.
Tapping Evaluate on a row starts the deep read and leaves you where you are. Go down the list tapping the ones you care about, then come back to the Evaluated tab. A role that's already been read won't be read again, whatever you press.
5 ·Applications
Open a role and you get its posting, what it's asking for, how it scored, and the buttons that build things: the gap analysis first, then a tailored resume, then a cover letter, then a review of how the pair reads to a screener.
The gap analysis is worth doing before the resume. It lists every stated requirement as met, partial or missing, and the resume is written inside those boundaries — so it will not claim the missing ones on your behalf.
What this won't do
It won't invent a claim to close a gap. It won't apply on your behalf. It won't tell you that you're a 9/10 fit when you aren't — a low score is the cheapest useful thing here, because it's the one that saves you an afternoon.
It also won't let you edit your record directly. Everything in it points at something you said in an interview, and a free-text box would quietly break that link. To change what's in there, say so in the interview.
What it costs
Reading one posting properly costs a few cents; a full set of documents about twelve. Your account page shows every run and what it cost, and there's a monthly ceiling that stops work rather than surprising you.