Does an ATS look at my design

No. It receives text extracted from the file. The design decides what that text looks like when it comes out, which is a different and more important thing.

Can I see what it reads

Something very close to it. The report here prints the words, lines and recognised headings from the extracted text, so a large gap between that and your page means text is being lost.

Is this checker an employer's ATS

No, and nothing public is. We have no access to those systems. This reads your file the way a parser does and reports what came back.

What is the most common failure

Reading order. A sidebar or a table can interleave two columns into one stream, so job titles arrive under somebody else's bullet points.

Do images and logos matter

They extract as nothing. A résumé whose content lives inside an image arrives empty, which is why a photographed or scanned page needs recognition first.

Guide · ATS

What an applicant tracking system actually reads

An applicant tracking system does not see your résumé. It receives the text pulled out of the file. Everything else follows from that.

ConxfolioPublished 2026-09-12

— it gets the text layer and nothing else.

An applicant tracking system does not look at a page. It takes the file, pulls the text layer out of it, and works on the string that comes back. Every piece of formatting advice worth following is a consequence of that one fact.

What extraction produces

A run of words in the order the file gives them up, which is not always the order your eye reads them in. A single-column document extracts in the order it is written. A two-column document may extract column by column, or line by line across both, depending on how the file was built and which library does the reading.

What gets lost

Page headers and footers. Often skipped, which is why contact details belong in the body of the page.
Text inside images. Extracts as nothing at all. A logo is fine; a résumé rendered as a picture is empty.
Icon glyphs. They come out as private-use characters or as replacement marks, and enough of them look like a damaged file.
Visual grouping. Whitespace, rules and boxes carry meaning to you and nothing to the parser. Headings are what survive.

Why headings carry the structure

Once the text is a single stream, the only way to tell where experience ends and education begins is a line that says Education. Systems keep lists of the headings they recognise. The checker here keeps one too, covering summary and its variants, experience and employment, education and qualifications, skills and competencies, projects, and a group for certifications, awards, publications, languages, volunteering and references. A heading outside the list is not treated as a section boundary, and the lines under it are read as part of the section above.

Dates, numbers and contact details

Dates are matched by pattern, which is why four-digit years in one format survive and shortened years do not. Email addresses are matched by shape. Phone numbers are matched by digit count and formatting, which is why an international number with a country code is safer than a bare run of digits.

What our reader adds, and what it refuses to do

When you import a file into the résumé builder, deterministic specialist readers run over the same extracted text to sort it into fields. Every string they produce is checked in code to be a span of your document, so a field that cannot be traced back to the source is dropped rather than guessed. Where they are uncertain they ask a question, and nothing is applied until you answer. No language model is involved at any point.

Checking your own file

Run the file you actually send through the checker and read the line that says what it read: the words, the lines and the recognised headings. Compare it with the page in front of you. A two-page résumé reporting sixty words has lost most of itself before any scoring happened, and no amount of keyword work repairs that. The format is the fix, and the format guide lists the five constructions that cause it.

What happens after the extraction

A system stores the text and usually tries to split it into fields: name, contact details, a list of roles with dates, an education block, a skills list. That is the same job the import in the résumé builder does, and it is where a scrambled reading order turns into visible damage, because a role can acquire somebody else's dates or a skills list can swallow a line of interests.

Recruiters then search that stored text. This is the part worth designing for: a document that splits cleanly into fields is a document that appears in the right searches with the right dates attached.

How much of this is guesswork

Some of it. Systems differ, versions differ, and no public page can tell you which one an employer runs or how it is configured. What does not vary is the first step: every one of them reads text out of a file. So the honest advice is the advice that is true regardless of the system, which is to make the extraction clean and the structure obvious, and to check the exported file rather than trusting the preview. The guide to the checker sets out exactly what can be measured from that extraction and what cannot.

See the extraction

The report prints the words, lines and recognised headings it read, so you can compare them with the page you think you sent.

Open the ATS checker See the 21 ATS-safe designs

Questions

Does an ATS look at my design: No. It receives text extracted from the file. The design decides what that text looks like when it comes out, which is a different and more important thing.
Can I see what it reads: Something very close to it. The report here prints the words, lines and recognised headings from the extracted text, so a large gap between that and your page means text is being lost.
Is this checker an employer's ATS: No, and nothing public is. We have no access to those systems. This reads your file the way a parser does and reports what came back.
What is the most common failure: Reading order. A sidebar or a table can interleave two columns into one stream, so job titles arrive under somebody else's bullet points.
Do images and logos matter: They extract as nothing. A résumé whose content lives inside an image arrives empty, which is why a photographed or scanned page needs recognition first.

Conxfolio is a free set of four career tools: a résumé builder with 37 rendered layouts, an ATS résumé checker that prints its own arithmetic, a cover letter builder that traces every proof paragraph back to the line it came from, and a portfolio builder with 20 authored designs. There is no account to create, nothing is held back for a paid plan, and no language model is used anywhere in the product, so the readers, the score and the letter are deterministic code you can check.