Runs entirely in your browser
Read your resume back the way software does
Drop in whatever you already have. It is read the way hiring software reads it, and you get told what came through, what did not, and what to change. PDF, Word, OpenDocument, rich text and plain text. Nothing is uploaded. The file never leaves this tab.
The report
What comes back, and in what order
Four panels, generated from your file rather than from a checklist somebody applied to it by eye. There is no preview of a paid version and no gate: the whole report is the free part, because the expensive half of this product is the editor and it is free too.
01
What was read
The format, the size, the page count, and whether the file carries a real text layer at all. If a hiring portal will not accept the extension, that is said above the number rather than inside it: a file the upload step refuses never reaches a parser, so how well it would have parsed is beside the point.
02
What a parser finds
Name, email, phone, links, section headings, employers, job titles and dates, each shown as recovered or missing, with a count. This is the panel most people came for. A field missing here is a field missing from the record a recruiter searches.
03
What is costing you points
Every finding, heaviest first, each labelled by severity and named specifically. Not a warning about fancy formatting: the layout table on page two, the heading a parser cannot map to a known section, the roles with no end date.
04
What now
The same document written back out cleanly in any of the export formats, or opened in the editor so you can fix what the report found and read the result back. Both run in the browser, on the text that was extracted, so what you send is what a parser saw.
One more panel sits beside the score and is deliberately outside it. Paste a job posting in and you get back which of its terms your resume actually uses. That answers a different question from whether your file can be read, so it does not move the number in either direction.
Formats
What you can check, and how far the check goes
How much a report can honestly claim depends on how much of the document survived into the file. A plain text file has no reading order to scramble, so it is never credited for keeping one. The depth is stated per format instead of implied.
| Format | Extension | What the report covers |
|---|---|---|
| Full report: reading order, columns, structure, fields and dates. | ||
| Word | .docx | Full report: reading order, columns, structure, fields and dates. |
| OpenDocument | .odt | Full report: reading order, columns, structure, fields and dates. |
| Rich text | .rtf | This format carries no page geometry, so column and reading-order checks are skipped. |
| Plain text | .txt | Plain text has no layout to lose. Fields, headings and dates are still checked. |
| Markdown | .md | Plain text has no layout to lose. Fields, headings and dates are still checked. |
Readable is not the same as submittable. Greenhouse publishes what it takes at the upload step, .docx, .pdf, .rtf, .txt, so OpenDocument and Markdown are flagged as a blocker before anything is scored, however cleanly they read. That warning is the one part of the report that is about the portal rather than about your document.
Privacy
Why there is no upload button
The file is read where you dropped it. A PDF is parsed in the page, a Word or OpenDocument file is a zip archive unpacked in memory, and the bytes are never put in a request body.
There is no server holding your employment history, which means there is nothing to breach, nothing to hand over, and nothing to quietly train a model on. That is a structural claim rather than a promise in a policy: the code is open, the network tab is empty, and you can check both.
It is also why there is no account. Documents you build here live in this browser’s own database. Clear the site data and they are gone everywhere, because this was the only copy, and nothing syncs between your laptop and your phone. Export the file if you want it somewhere else. That trade is the price of the paragraph above, and it is worth stating rather than burying.
Questions
Reasonable things to ask
Collapsed, because most people want one of these rather than all six.
Will an ATS reject my resume for formatting?
Usually not directly. What happens is worse in a quiet way: the file is parsed badly, and a recruiter sees a record with fields missing. You are competing against candidates whose records came through complete.
PDF or Word?
Send what the posting asks for. If it does not say, PDF is safer because it carries its own fonts and cannot reflow. Both export cleanly here, and both are checked the same way.
Is one page still the rule?
Under roughly ten years of experience, yes in the US and UK. Beyond that two pages is normal. In Germany and much of Asia longer is expected. This is convention, not a parser constraint, so it is flagged as a suggestion.
Do keywords from the job posting help?
Matching real terms helps, since recruiters search on them. Stuffing does not, and it is obvious to a human reader and to modern matching, and it costs you the interview.
What happens if I clear my browser?
Your resume is gone, because it was only ever stored on your machine. Export the JSON if you want a backup. That is the trade for never uploading anything.
How will you make money?
Currently there is no plan to. The core will stay free. If that ever changes it will not be by paywalling the download, because that is the thing this exists to avoid.
After the score
A number on its own is not advice
The report names what broke and ranks it. Two pages carry the rest.
If you want to know why a given choice breaks a parser, the guide walks the four stages every applicant tracking system puts a file through and shows where each one loses data. If you want the arithmetic behind the number, the methodology page publishes every weight, its source, and which of them are our judgement rather than anyone’s research. No other checker in this category does the second one, which is most of why this one exists.