What can a parser actually recover?
Paste the text and see what comes out. Contact details, headings, dates and bullets — checked against the things that genuinely break extraction, with no claims about anyone's applicant-tracking system.
What a parser can recover from your resume
Paste the text and this checks the things that are actually checkable: contact details, headings, dates, bullets, and the formatting constructs known to interfere with extraction.
Which of these does your resume use?
Copying the text cannot reveal layout, so these are yours to tell us. Each is a risk, not a verdict.
Findings
What was recovered
How the score is built
Seven dimensions, weighted: parseability 30, formatting risk 20, section recognition 15, contact recovery 10, chronology 10, evidence strength 10, readability 5.
A dimension that does not apply is dropped and the rest renormalised, rather than scored as zero. Certifications are not a defect on a resume for a role that never asked for one.
Critical failures cap the result. A document whose text cannot be recovered is not a 62 because everything else was tidy. The caps are: no recoverable text 30, contact lost in export 55, material content lost 60, reading order broken 65.
Pasted text cannot measure export fidelity. Round-trip parseability has to be tested on the actual generated file, so from pasted text this is assessed on what the text itself shows and the score is presented accordingly.
Why nobody can tell you if you will "pass"
Applicant-tracking systems are products bought and configured by employers. The same product behaves differently at two companies because the screening rules, required fields and knockout questions are set by whoever configured it.
So any tool claiming a pass rate, a compatibility percentage, or approval by a named system is describing something it cannot see. What can be tested honestly is the document: whether the text comes out, whether the sections are recognisable, whether the dates are unambiguous.
The failures worth caring about
Almost every genuine parsing disaster is one of a small set: contact details in a header or footer, text inside an image, content in a floating text box, or a two-column layout read across instead of down.
Each loses information that no amount of keyword optimisation recovers. A phone number in a footer is not a formatting preference — on several extraction paths it is simply absent, and the application arrives with no way to contact you.
Ambiguous dates cost more than they look
"03/04/24" is March 4th to an American reader and 3rd April elsewhere, and no parser can resolve it from the string. "Mar 2024" reads identically everywhere and takes the same space.
This is the cheapest fix on any resume and one of the most common problems.
What we will not do
This will not write a bullet for you, invent a metric, or add a qualification you did not mention. Where something is missing it says so and asks what you actually did.
That is not caution for its own sake. A resume is a document you will be interviewed against, and a tool that quietly improves it with things that did not happen has set you up to be found out in the room.