What an ATS needs from a CV
An applicant tracking system stores each CV and usually extracts its text into a searchable profile: name, contact details, employers, dates, skills. Whether that works depends less on design than on structure. If the text is real, in the main body and in a sensible order, almost any parser copes. If it is an image, split across text boxes, locked in a header or spread over two columns, the extracted profile can come out empty, jumbled or missing contact details, and a recruiter searching the database may never find the candidate.
There is no single ATS standard and vendors do not publish their parsers, so no free tool can promise how one specific system will behave. This checker reports the structural signals that most often cause problems, explains each one and shows you the extracted text so you can judge for yourself.
The checks and why they matter
- Text layer: a scanned or photographed CV has no selectable text. Without optical character recognition, a parser sees nothing.
- Pages with little or no text: often a scanned page, a page built from images or text converted to outlines by design software.
- Page count: measured from the text pages of a PDF, estimated for Word files.
- Tables, text boxes and columns: layout tables and sidebars can be read across rather than down, mixing a skills list into a job history.
- Headers and footers: some parsers skip them, so contact details that live only there can be lost.
- Unusual characters: icon fonts for phone and email symbols, ligatures such as “fi”, replacement marks and invisible characters can break search.
- Contact details in the body: an email and phone number written as plain text in the main body are the most reliable.
- File size: large files usually mean embedded images or fonts.
Read the reading order
The most useful single check is to look at the text a parser actually gets. The preview shows the first part of the extracted text in the order it comes out. A clean CV starts with the name, then contact details, then a heading and its content. If you see a skills sidebar woven into a job description, or the contact line at the end, the layout is working against the candidate.
What this checker does not inspect
It does not list fonts, measure images, check colour contrast or render the page, and for PDFs it estimates columns from the order of the text rather than measuring page geometry. It does not judge content: whether the CV fits a job is a different question, answered by the CV keyword checker and, ultimately, by a recruiter reading it.
Fixes that usually help
- Export from the original document instead of scanning or printing to image.
- Use one column, standard headings (Profile, Experience, Education, Skills) and a common font.
- Write contact details as words in the body: “Email: name@example.com”, not an envelope icon in a header.
- Move text out of text boxes and layout tables; accept all tracked changes.
- Keep dates and job titles on the same line or directly under the employer.
How your file is handled
The file is uploaded once, without cookies, to our file reader. It is parsed in memory by the same isolated document service that reads CVs in the app, reduced to counts and flags plus the short reading-order preview you see, and discarded. Nothing is written to storage or logs. Files are limited to 5 MB, 20 pages and a few checks per visitor every few minutes.
For recruiters sending CVs to clients
Clients load your submissions into their own systems. CVPitch exports each approved CV as an editable DOCX and a matching text-based PDF in your agency’s branding, with contact details handled by your policy. See pricing or try 10 CVs free.