How to Check Whether Your Resume PDF Is ATS-Readable
A repeatable seven-step test on your own PDF: copy, paste, read the order, check the fields. What a passing result proves, and what it does not.
Copy the text from your PDF into a plain-text editor and read the result. This can reveal missing content, broken characters, and a confusing reading order before you submit an application. It shows how your PDF viewer handles the file; an employer’s parser may behave differently.
The checks below cover extraction in your viewer and, when available, the fields populated by the application form.
Step 0: Check what the employer asked for
Before testing the file, check whether it is the right kind of file. Application forms usually list accepted types, and that list overrides every general rule. Greenhouse, for example, accepts .doc, .docx, .pdf, .rtf and .txt with a 100 MB limit (checked 6 September 2026), but that is one vendor’s list, and yours may be narrower.
Expected result: the posting either names a format, in which case you are done deciding, or says nothing, in which case the PDF-versus-Word trade-off applies.
Step 1: Can you select the text?
Open the PDF in any viewer. Click into the middle of a bullet and drag across a few words.
Expected result: the words highlight, character by character.
If nothing highlights, check whether the file is image-only or the viewer restricts selection. Try a second viewer or export a fresh text-based PDF from the source document. OCR can recognize text in an image, but you should not rely on the employer’s system to do that. A viewer may also perform recognition itself: Apple documents copying image text with Live Text. Selection alone therefore does not prove an original text layer exists.
Step 2: Select all, copy, paste into plain text
Cmd/Ctrl+A in the PDF, copy, and paste into a plain text editor — Notepad, TextEdit in plain mode, or any code editor. Not Word, which will re-render some of what you are trying to inspect.
Expected result: a plain, ugly, complete version of your resume.
This paste is the copy you will use for the next four steps. Keep it open.
Step 3: Read the order
Read the pasted text top to bottom.
Expected result: the same order as the page — header, then summary, then each role with its bullets underneath it, then education, then skills.
Signs of a defect:
- Interleaving. A line from the left column, then a line from the right, alternating. This is the classic two-column failure, and it happens because the extraction runs across the page rather than down a column.
- Bullets separated from their role. All the job titles first, then all the bullets in a block.
- Text appearing at the wrong end. A header or footer that lands in the middle.
A two-column layout is not automatically broken. Inspect the pasted text to see how this viewer orders the columns. Another parser may use a different order.
Step 4: Find your contact details
Search the pasted text for your email address and phone number.
Expected result: both present, intact, and near the top.
Signs of a defect: an email split across lines, a phone number with characters dropped, or contact details missing entirely. Check whether the details sit in a header, footer, text box, or image that the viewer did not extract. Put the contact text in the main body near the top and try again.
Step 5: Look at the bullets and special characters
Scan the pasted text for the characters that travel badly.
Expected result: bullets appear as a consistent marker or as plain line breaks; accented characters, dashes and quotation marks appear as themselves.
Signs of a defect: é where an é should be, boxes or question marks, or a bullet glyph that arrived as an unrelated symbol. Possible causes include font mapping and encoding. Try ordinary bullet markers and check whether the new export preserves accented names and punctuation.
Step 6: Check the section headings survived
Search the pasted text for the words Experience, Education and Skills.
Expected result: each appears as its own line.
Signs of a defect: a heading merged into the paragraph after it, or one that has been styled so heavily it extracted as separate letters. Note too that unconventional heading names — Where I’ve Worked, My Journey — extract perfectly well but give field detection less to go on. The general structural rules are covered in what makes a resume ATS-friendly.
Step 7: Upload it and read the form
The most informative test is the one the employer’s own system runs. Start an application at the company you are applying to, upload the file, and stop before submitting.
Expected result: the fields the form pre-fills — name, email, phone, most recent employer, dates — contain the right values.
These fields show the result of this application form’s processing. Some forms populate only a few fields or none at all. An incorrect value identifies something to correct; an empty form alone does not tell you whether parsing ran.
Correct the fields by hand before you submit. A parse failure is not a rejection. Greenhouse’s own documentation describes what happens when parsing fails: the resume “has only been attached to the candidate,” and the details are entered manually (Greenhouse support, checked 6 September 2026). Check what this form allows, correct the values, and complete the application when you are ready. Uploading a file can transmit it to the employer even before final submission, so use this step for applications you intend to make.
Fixing what you found, then re-testing
For visible layout problems, check the relevant formatting settings, including spacing and page breaks. For missing or corrupted text, check the source document and export format as well.
If you rebuild the file, re-run steps 1 through 3 on the new download rather than assuming the fix carried. Running this protocol costs nothing whatever built your PDF — the test is you and a text editor. In ExportMyResume, exporting a replacement PDF requires Pro, as does Word. Free for now during the launch offer: sign in to download PDF or Word without a subscription. Plain text is also free, and signing in enables PNG without a watermark. A PNG is still an image that would need OCR for text recognition.
What a pass actually proves
A clean result shows that your viewer can produce readable text in a sensible order and that the details you checked are present. Checking the application fields adds evidence about that form’s processing. Neither result proves compatibility with every system.
After a repair, repeat the checks on the new download. Keep the version you actually inspected as the one you attach to the application.