See what comes out of your file
Upload the file you actually send to employers. This extracts it the way a parser would and shows the result — recovered text, the fields it could find, and what it could not.
What comes out of your file
Upload the actual file you send to employers. This extracts it the way a parser would and shows you the result — the recovered text, the fields it could find, and what it could not.
Drop your resume here, or click to choose.
PDF, DOCX or TXT. It is read in this tab and never uploaded.
What this is actually doing
A PDF is read with pdf.js, which is the same engine Firefox uses to display one. A DOCX is unzipped and the text of its document body read directly. Neither is an employer's parser, and no employer publishes theirs.
What they have in common with any parser is the hard part: a PDF stores glyphs at coordinates, not paragraphs, so reading order has to be reconstructed. Where a tool of this kind scrambles your resume, so will most others.
Fields are found by pattern, not by understanding. An email is an email; a name is a line near the top that looks like a name. Where the guess is wrong you will see it, which is the point of showing it.
Both libraries are fetched from a public code CDN the first time you use the page, before you choose a file and carrying nothing about you. Your file goes nowhere.
The file is the thing being read, not the design
A resume that looks correct in Word can come out of extraction with its columns interleaved, its headings indistinguishable from body text, or its contact block missing entirely. None of that is visible until something reads it.
Two constructs cause most of it: multiple columns, and text placed in a header, a footer or a text box. All of them render perfectly and extract badly.
If your name is not in the recovered text, nothing else matters
Contact details in a header are the most common way a resume arrives at an employer anonymous. The fix takes ten seconds — move them into the body of the document — and it is the highest-value change on this page.
A scan cannot be read at all
If you cannot select the text in your own PDF viewer, there is no text in the file, only a picture of it. No parser recovers anything from that, and optical character recognition is a different job with a much higher error rate. Export a real PDF from the original document instead.