What does an ATS parsing test actually check?
Applicant tracking systems don't read resumes the way humans do. They extract the document into a single text stream and feed that stream into a parser that classifies sections, pulls out fields, and indexes keywords. When the extraction is clean, your resume is searchable and ranked correctly. When the extraction is broken, your strongest achievements may simply not exist in the system.
The most common failure points are layout-driven: multi-column templates, decorative icons, text-boxes, scanned PDFs, and tables used for layout instead of data. All of these look fine on screen but scramble when an ATS runs its text extractor. The parsing test catches them by running the same extraction step and surfacing the artifacts.
Why deterministic — not an "AI score"?
A percentage that changes each time you re-upload the same file isn't an assessment — it's a guess. The parser test uses fixed heuristics documented in lib/parser/diagnostics.js: same input, same output, every time. You can verify every finding directly in the extracted-text panel.
How to make your resume ATS-readable
- Single column. Sidebars and two-column templates extract scrambled.
- Plain ASCII bullets. Custom glyph icons from fancy templates often vanish during parsing.
- Standard section headers. Use the exact words "Experience", "Education", "Skills" — not invented names.
- Header text in the body. Avoid placing name + contact info inside Word's header element — it can land below the body in the extracted stream.
- No tables for layout. Tables get flattened. Inline dates and titles ("Acme Corp — Sales Manager — 2020–2023") parse cleanly.
- Text PDFs, not scans. A scanned PDF extracts as zero text. Re-export from Word or Google Docs.
Frequently asked questions
What is the ATS Parsing Test?
It's a transparent diagnostic that runs the same kind of text extraction an applicant tracking system would, then surfaces formatting and structure issues that commonly break parsing. It is not an AI score generator and it does not invent a percentage out of thin air.
How is this different from an 'ATS score'?
An 'ATS score' usually compresses a resume into one fake number. The parsing test shows you the actual extracted text, points at the specific lines that are likely to break, and explains why. Every finding cites the line it comes from so you can verify it yourself.
Which ATS systems does this emulate?
None specifically. ATS vendors don't publish their parsers. What we do is run deterministic heuristics that catch the same patterns those parsers fail on — multi-column layouts, missing section headers, corrupted bullet glyphs, scanned PDFs, header/footer contamination, and so on.
Will my resume be stored?
No. The file is parsed in memory and discarded. Only if you choose 'Copy report link' is an anonymized diagnostic snapshot cached for 7 days at a unique URL.
What file types are supported?
PDF, DOCX, and plain text up to 4MB. You can also paste resume text directly. PDFs are the most informative test, since that's what most ATS systems actually ingest.
Does pasting text give the same result as uploading?
Paste-mode skips PDF extraction entirely, so it can't detect column-merge or symbol-corruption issues that come from a PDF specifically. Upload the original file if you want the full picture.
Why does my resume look fine to me but fail the test?
ATS systems don't see what you see. A two-column resume looks tidy in Word but extracts as a scrambled single text stream. The parsing test shows you that stream, side-by-side with your formatting findings, so the gap is obvious.