VULTURE

A free AI check for résumés and cover letters

See how much of your writing reads like a language model, which lines give it away, and what to say instead. It runs in your browser against a rule set, never by asking another model, so nothing you wrote is uploaded anywhere.

Check my résumé freeEvery feature

What it looks for

No model is asked for an opinion. The check is a rule set built from the tells that large language models leave behind, each one carrying the evidence that triggered it:

VocabularyThe words whose use jumped after 2022: delve, leverage, robust, seamless, spearheaded, meticulous, showcase, foster. Each flagged word comes with a plainer one to use instead.
Stock phrasing"Not only... but also", "it is worth noting", "in today's fast-paced world", the tricolon, and the negative parallelism that models reach for when a sentence needs weight it has not earned.
RhythmSentences of almost identical length are the strongest structural tell. Human writing is uneven; model writing is metronomic.
Formatting habitsBold lead-ins on every bullet, em dashes in a résumé, participial tails ("...improving efficiency by streamlining processes") and lists of exactly three.
Missing human markersContractions, specific numbers, named tools, odd details. Their absence counts against the text as much as any flagged word counts for it.
Model fingerprintsPhrasing particular to one family of models, so the report can say which house style the text reads like.

An honest word about AI detection

No detector, ours included, can prove who wrote a text, and any tool that claims otherwise is selling something. What this one gives you is evidence: which lines read as machine output and why. That is the useful question anyway. A recruiter is not running a detector; they are reading four hundred résumés that all sound the same, and yours has to sound like a person who did the work.

Résumés also score higher than ordinary prose by nature: they are formal, list-shaped and stripped of contractions. Read the number next to the flagged lines rather than on its own.

Fix it, do not just measure it

Inside the free app, every flagged line comes with the reason and a rewrite in plain English, with your own choice of English, tone and length. You can also check any text, not only a résumé: a cover letter, an outreach message or an email to a recruiter.

It is part of Vulture, a free, local-first job hunt app by Harsh Sharma and Gixts Labs, with a résumé builder, ATS scoring, a job search and an application tracker. Nothing you write leaves your device unless you ask it to.

Questions

Is the AI detector free?
Yes. The score costs nothing and needs no account, because it runs inside your browser rather than on a server. The full report, with the flagged lines and the rewrites, is in the free app.
Does my text get sent anywhere?
No. The check is a rule set that runs in this tab, and it never calls a model. Nothing is uploaded and nothing is stored.
How accurate is it?
Treat it as evidence, not a verdict. No detector can prove authorship. What it can do reliably is point at the sentences that read like default model output, which is the part you can act on.
Why does my own writing score high?
Résumés are formal, list-shaped and short on contractions, which is also how models write. That is why the report shows the signals: if the score is high but the flagged lines are few, the writing is probably fine.
Can it tell which AI wrote something?
It can say which house style the text reads like, based on phrasing that is particular to one family of models. That is a hint, not an identification.
Does using AI on my résumé hurt my chances?
Using it as a first draft does not. Sending the first draft does. Recruiters recognise the pattern, and the tracking system rewards specifics the model cannot invent: your numbers, your tools, your outcomes.
Score a résumé nowGet the desktop app