The scanner
Your resume isn’t rejected by a person. It’s rejected by software.
Before a recruiter reads a single line, an applicant tracking system parses your file into fields. Get one date wrong, bury a title in a text box, write a bullet with no number — and you’re filtered out with no reply, no reason, no second chance. ProofKit runs the same class of checks that software runs, shows you every failure in plain English, and tells you the exact fix. Free, in your browser, in seconds.
No account needed — try the full 23-check scan now. About four seconds.
How it works
Three steps. No black box.
Every score comes with the exact checks behind it. You never pay to find out what was wrong.
Paste or build
Type your resume in the builder, or import a file. Nothing leaves your browser until you ask for a scan.
Scan against 23 checks
Contact parsing, section structure, bullet quality, dates, length, verbs, formatting, and job-description keywords — the same rubric, every time.
Fix what it flags
Each failed check names the problem and the fix. Re-scan until it’s clean, then export and share a proof of the score.
The rubric
All 23 checks, on the table.
No mystery score. These are the exact categories the scanner runs — the same list your export is scored against.
Contact (4)
- Name present — parsers look for it first
- Email parses — plain text, no images
- Phone parses — one standard format
- Location parses — city and country
Structure (4)
- Summary 40+ chars — two lines beat zero
- Experience section — the section every parser hunts for
- Education section — degrees, schools, years
- Skills 5+ named — matched against the job later
Bullets (4)
- Bullets present — paragraphs get skimmed and skipped
- Quantified 60%+ — numbers are what recruiters quote
- Strong openers 5+ built, led, cut, migrated
- 28 words or less — long bullets lose the reader
Dates & length (3)
- Dates with years — gaps get questioned, missing dates get rejected
- Length ~1 page — one search, one story
- No I / me / my — resumes drop the pronoun
Skills & verbs (3)
- Action verbs 5+ — managed is not a result
- Hard skills named — the tools you actually used
- No duplicate bullets — copy-paste reads as padding
Format safety (3)
- No tables or text boxes — the #1 silent parse killer
- ASCII-safe characters — smart quotes break fields
- No photo dependency — many parsers discard images
JD targeting (2)
- JD keywords matched — paste the job post, see what’s missing
- Title mirrors the role — your headline should match theirs
Parser compatibility
Four ATS behaviors, modeled. Twenty real documents, tested.
Every competitor shows you a score. This is the part they don’t: the rules the score comes from, and a test run you can watch. These are modeled parsing behaviors, not vendor certification — we can’t run Workday, Greenhouse, Lever or iCIMS locally, and every employer configures its own parser. What we can do is test the document shape our exports produce against behaviors we can name, and show you exactly where a table, a text box or a two-column layout breaks.
Running the corpus…
See the 20 test documents and what each one breaks
Questions
What people ask before their first scan.
Is the scan really free?
Do you store my resume?
Why did my score drop after a “fix”?
Does a 91 mean I’ll get interviews?
Find out what the parser sees.
One scan. 23 checks. The fixes, named. Then export a clean file and a link that proves the score.