How it works
Four things happen in the seconds after you submit the form — in this order, every time. Nothing here is hidden inside the model; it’s ordinary code you can read in the repository.
One click fires four independent calls to Claude in parallel, each reading the same source documents:
.docx or copies as Markdown, built from the same structured data you see on screen — not a second pass through the model.The truthfulness rule
Every prompt is instructed to rephrase, reorder, and emphasize only what’s already in your resume and LinkedIn profile — never to add an employer, title, date, degree, tool, or metric that isn’t already there. That instruction is backed by a second, independent check: ordinary code diffs the output against your source documents and flags anything it can’t verify. Both show up in the margin next to your tailored resume:
Something the posting wants that your resume doesn't show. Named, not filled in — you decide what to do with it.
A term in the output that doesn't appear in your resume or the job posting — the deterministic check flagged it, so it's worth a second look before you use it.
Privacy
This tool is stateless: there’s no login, no account, and no database. Your job posting, resume, and LinkedIn content exist only in memory for the duration of your request, sent directly to Anthropic’s API to generate your results, then discarded. Reloading the page clears everything. Nothing you submit is written to a log, on this server or anywhere else in this codebase.