Generic AI resume writers fill gaps with fiction: a rounder number here, a bigger title there. Ours cannot. It tunes what you already wrote, then a separate pass checks every line against your own history. If a number, title, or employer is not already there, it does not appear. Then it shows you the full teardown: every change, next to your original, with the reason and the exact line in your record it came from. That is the point. You never take our word for what changed.
The rewriter reads the actual job description and re-orders, re-emphasizes, and re-words the bullets already in your history. Nothing is added from outside your record.
A separate, independent pass compares each rewritten line to the resume you gave us. Any number, title, employer, or credential not already present gets flagged, and the rewrite is discarded in favour of your exact original words.
The annotated view puts your original and the tuned version side by side. Every change carries a plain reason (“used the words this job's screener expects”) and the exact quote from your own record that backs it. The gaps we could not honestly fill are named, not faked.
Used the words this job’s screener expects, and led with ownership.
From your own record
“worked with department heads to resolve variances”
Mentioned in skillsShown in your workThe check covers hard facts: numbers, dates, employers, titles, and named skills. It does not yet fully judge tone. The line between “contributed to” and “led,” when both are true, is handled by a tightly limited prompt, not the same hard check. We are building stronger checks on meaning; until they ship, read a tuned bullet's tone as our best editorial call, not a guarantee. We would rather show you exactly where the guarantee ends than let the word “checked” cover more than it does.
Not this one. Every generated line is checked by code against your own resume and rejected if it cannot be traced to something you actually logged. If a claim cannot be verified, we keep your original sentence untouched rather than soften it into a plausible guess. The tuning is real (re-ordering and re-emphasis of your own record) but the facts inside it are never new.
It shows you. After tailoring, an annotated before/after view puts your original resume next to the tuned version, and every single change carries a plain-language reason and the exact quote from your own record that supports it. Nothing is changed silently. You can read the full teardown before you download anything.
There is no rigorous study showing that keyword tailoring by itself raises interview rates, and we will not pretend otherwise. What is measured: a randomized experiment with about 500,000 jobseekers found that better resume writing raised hiring probability by 8% [van Inwegen, Munyikwa and Horton, 2023]. That study measured writing quality, not keyword tailoring, so treat tailoring as making your real experience easier for a screen to find and read, not as a guaranteed odds boost.
It stays honest instead of getting bigger. If a job description calls for something your history does not show, the rewrite will not manufacture it. You will see the closest real bullet re-emphasized around what you actually have, or, if there is nothing to re-emphasize, your original wording kept as is, flagged as an honest gap. A thin match is real information; inventing a match is just a more convincing way to fail the same screen later.
Finding jobs and scoring your fit are free. Bring your resume, and we tune it to the job using only what is already true about you.