Why I made a resume builder read its own PDFs

I built ATSPrep after getting tired of resume tools returning a score without a receipt. A 78 or 92 feels useful until you ask: what exactly did the parser read, and why is this bullet considered weak?

So I started with a simple rule: ATSPrep should show its work.

Paste a job description and the matching evidence appears inside the resume-not in a disconnected keyword report. Weak bullets are flagged for what is actually missing: a number, scope, or outcome.

The part I care about most happens after export. X-Ray opens the generated PDF, extracts the text again, and shows what survived. If a tool calls a resume “ATS-friendly,” I think it should demonstrate that.

Everything runs locally because resumes contain information I would not want sitting in another startup’s database.

ATSPrep is free, requires no account, includes seven scored role templates, and exports PDF and DOCX.

If you try it, I’d appreciate one honest answer: where does the workflow become confusing or make you hesitate?