How it works
The mental model: your profile is source code, and each application page is a build target. You maintain one evidence base; every page is compiled from it for one specific job.
- 01
Import what you've already written
Upload a resume. We pull it apart into structured evidence — roles, projects, numbers, the story behind each one — rather than keeping it as a document. This is the part that compounds: every application after the first draws on the same base.
- 02
Paste a posting
We read the posting, research the company for a few minutes, and show you what we understood before anything is built on it. Then we pick which of your evidence answers which requirement.
- 03
Answer only where there's a gap
Where a requirement isn't covered, you're asked — up to five questions, each with the reason it's being asked. Answers become permanent evidence, so the next application in the same territory asks less. Saying "I haven't done that" is a real answer: the claim stays off the page.
- 04
Review the page
You get a structured page, not a blob of generated HTML: reorder sections, edit any line, dial the company's branding up or down. A provenance panel shows which evidence every claim came from, and the page won't publish with a claim that traces to nothing.
- 05
Send it, then find out
Publishing mints a URL that exists nowhere else. When something opens it, we score how confident we are that it was a person — and tell you plainly when it was a scanner instead.
About the AI, specifically
The model never writes freeform prose into a page. It fills a schema — headline here, three requirement cards there — and a component system renders it. That is why generated pages look designed rather than generated, and why they can’t drift into something you wouldn’t send.
It also cannot invent. Every claim must cite a piece of evidence you provided, and the pipeline refuses to publish a page containing a claim that traces to nothing. If your experience doesn’t cover a requirement, the page says so or stays quiet — it does not make something up to fill the gap.
Questions worth asking first
- What does the recruiter actually see?
- A page at an address like
yourname.suited.page/k7m2x9qp, in the company’s colours, with a footer that reads: “A personal application page by [you]. Created for a specific job application to [company]. Not affiliated with, endorsed by, or published by [company].” The footer cannot be removed. There is a page written for them explaining what they’re looking at and what was recorded when they opened it. - Can my current employer find this?
- Not by looking. Pages are never indexed, there is no directory, and the address is unguessable. Only someone you send the link to can open it. The people who can see your applications are you and whoever you send each one to.
- What if the ATS strips the link?
- You also get a complete, conventional one-page resume as a PDF — plain formatting, standard fonts, single column — derived from the same page so the two can’t disagree. Upload that where a file is required; put the link in the header, the cover note, or the email.
- How long does it take, and why?
- Importing a resume takes a minute or two. Researching a posting takes five or six, because it actually reads about the company rather than pattern-matching the job title. Generating the draft takes two or three. You review it in about three. A first page in twenty-five minutes; later ones faster, because your evidence is already there.
- Is this only for tech roles?
- No. The page’s section mix and tone follow your field: a clinical role leads with credentials where an engineering role leads with projects, and a hospitality page reads differently from a marketing one. Six exemplars span those fields.
- Does the AI make things up?
- It can’t publish anything it can’t trace to evidence you gave it — see “About the AI” above. Whether that evidence is true is on you, as it would be on a resume.
See it on a real posting
Bring one you actually care about.
Request an invite