Why us
How we score
28 fixed rules, each with a weight and a tier. No AI in the number.
Source, 28 rules, 13 Oct 2026, changelog
The rules
Your score is the weighted average of 28 rules. Each rule checks one thing: a section is present, the contact details parse, a bullet has a number. It returns a score from 0 to 100, a one-line reason, and a fix.
A rule at 80 or more reads ok. From 50 to 79 it reads warn. Below 50 it reads fail.
Weights are fixed whole numbers from 3 to 12. The same resume always gets the same score. The full table, with your score on every rule, opens from any report under “How we scored your resume”.
- Required sectionsweight 12
- Contact informationweight 10
- Quantified impactweight 10
- Grade formatweight 3
Source, rule registry, 28 rules
Four tiers
A tier is the rule's category. Ok, warn and fail are this resume's result. A tier 1 rule at ok means we checked the killer and it is fine.
Tier 1: Resume-killer
11 rulesAn immediate filter. Fix these before anything else.
Tier 2: Strong-impact
10 rulesCosts interviews. The second pass.
Tier 3: Polish
6 rulesCompounds over time. Safe to defer.
Tier 4: Recruiter credibility
1 ruleA human-credibility signal, not an ATS filter: a skill listed but never shown in a bullet.
Source, rule registry, 28 rules in four tiers
Rules we could not measure
A rule with nothing to assess is not scored. A bullet-quality rule on a resume with no bullets, or an email-format rule with no email, is left out.
Its weight moves to the rule that reports the gap. “Add bullet points” reports the missing bullets once and carries their weight.
This closes a loophole. Without the transfer, deleting weak bullets would shrink the denominator and raise the score. Measured before the change: deleting bullets took a resume from 74 to 77. After it: from 73 to 52.
Your report lists not-scored rules separately, with the rule that took their weight, and every sum on it uses the weight the engine used.
Source, engine notes, effective weight shown on every report
Three pillars
Beside the flat score, the report shows three pillars: JD match 40%, Parseability 40%, Content quality 20%. That split is our reading of what matters for an early-career resume. No ATS vendor publishes a scoring formula, so we do not claim one.
Without a job description the two remaining pillars renormalise to 67% and 33%.
Source, our weighting, not a vendor's, the evidence behind the rules
What to fix first
The top fixes on a report are ranked in this order: resume-killers first, by tier; then by the points you would recover; then fail before warn.
The points for a fix are the rule's share of the total weight, times how far the rule is short of 100. Rules already at full marks, and rules we could not measure, are never listed.
Source, engine notes, top fixes ranking
What the AI may do, and may not
May
- Rewrite a bullet you choose.
- Draft a cover letter.
- Tailor a version to a job description.
- Read a job description for its requirements.
- Suggest changes to your LinkedIn text.
- Write interview questions from your resume and the job description.
May not
- Change the score.
- Add a skill your original does not show.
- Change your names, employers, dates or education.
- See your email, phone, links or name. They are replaced before the text leaves.
- Run without you pressing the button.
Each AI action costs a credit, and the cost is shown before you press. A number the AI adds is flagged for you to confirm. The providers are named on the privacy page, as providers and never as models.
Source, our promise, privacy, security
Version
What changed in the rules, and when, is in the changelog. Why these rules and not others is on the evidence page.
Source, 28 rules, 13 Oct 2026, changelog