HomeAI Detector › For CVs and Resumes

AI Detector for CVs and resumes, built for job seekers writing their own authentic voice.

Scan CV bullets, resume summaries, cover letters, and LinkedIn About sections before ATS systems and hiring managers see them. Sentence-level highlights flag the ChatGPT-rewritten lines, so you can replace them in your own voice and keep the polish underneath. Built for solo job seekers running a search, not for HR teams. Free to try. No card.

Start free, no card See pricing
3 scans/day free Sentence-level highlights No training on your writing
Who it is for

Built for CVs, resumes, cover letters, and LinkedIn profiles.

Solo job seekers running a search across two-page corporate resumes, longer academic and executive CVs, cover letters, LinkedIn About sections, and portfolio descriptions. One person doing the writing, often a fresh search after a layoff or a planned move, mixing AI assistance with their own voice.

A job search in 2026 is a writing project as much as it is a networking project. The resume gets read by Workday or Greenhouse before any human sees it, the cover letter gets opened by a hiring manager who already has thirty other applications in the queue, and the LinkedIn profile gets skimmed by a recruiter inside a saved-search alert. Every surface is short, every surface is read fast, and every surface now competes against candidates who use the same AI tools you do.

CV and resume bullets

The two-page corporate resume and the longer academic or executive CV share the same writing risk: bullets that all read alike, summaries that lift the same opening line, and accomplishment statements that compress real work into ChatGPT phrasing. Scan the document and the highlights show which bullets to rewrite first.

Cover letters and LinkedIn About sections

Cover letters are the most personal genre in the packet and where AI flavour hurts the most. LinkedIn About sections are the second most read surface in a recruiter pipeline and the one most candidates leave on autopilot. Both benefit from the same workflow: scan, rewrite the flagged sentences, keep the personal voice underneath.

Portfolio descriptions and project write-ups

Designers, product managers, engineers, and writers all run portfolios that include short prose descriptions of each project. Hiring managers read these descriptions before opening the work. Generic AI-flavoured project blurbs reduce the click-through rate to the actual work, and the writing gets judged before the work does.

The ATS reality

Workday, Greenhouse, Lever, and iCIMS now flag AI-rewritten bullets.

Through 2025 the major applicant tracking systems added AI-content signals to their candidate review layers. The flag does not always auto-reject, but it does land the application in a lower review tier where a recruiter spends less time on the document before triaging.

What changed in the ATS layer

Workday added a writing-quality flag on the recruiter view in early 2025 and expanded it through the year. Greenhouse and Lever both surface third-party AI scores through their integration partners, which most enterprise customers have switched on. iCIMS does similar at the enterprise tier. The flag does not always mean a rejection, but it does mean the application enters a lower review tier where a recruiter spends ninety seconds instead of three minutes on the document.

The signal types ATS look at

ATS-side detection looks at sentence rhythm, action-verb vocabulary, and the templated opening summary that ChatGPT defaults to. The signals overlap with what hiring managers themselves notice once they have read enough applications, so the ATS flag is increasingly aligned with the human read. Pre-scanning your bullets and rewriting the flagged ones is the workflow that keeps you out of the lower tier.

The senior-candidate effect

Senior candidates and executives applying for VP and above roles get a closer human read, which means the templated-summary opening is even more visible in that pool. Retained search firms screening executive CVs spend ten to twenty minutes per document and notice voice issues fast. The fix is the same on every level of seniority: scan, rewrite the flagged sentences, keep the actual experience underneath.

Plans & pricing

Solo job seekers usually run on Free or Starter.

Most job seekers complete a full search on Free or Starter. Pro is the right fit if you are also rewriting LinkedIn and a stack of cover letters across a long search. Business is built for hiring teams, not for individuals. Full details on the pricing page.

Free
$0/forever

 

Casual search of one or two applications a week.
  • 3 scans / day
  • 5,000 chars per scan
  • Sentence-level highlights
  • 2 lifetime AI rewriter uses
Start free
Pro
$14.99/month

Billed $179.88/year — Save $60

Long searches with cover letters and a LinkedIn rewrite.
  • Unlimited scans
  • 50,000 AI rewriter words/mo
  • 10,000 chars per scan
  • 90-day scan history
Get Pro
Business
$29.99/month

Billed $359.88/year — Save $120

Built for hiring teams, not for individual job seekers.
  • 100,000 AI rewriter words/mo
  • 5 team seats
  • REST API access
  • White-label PDFs
Get Business

Yearly billing saves 25%. View full pricing →

Document genres

Five surfaces inside one job search.

CV and resume bullets, summary paragraphs, cover letters, LinkedIn About sections, and portfolio descriptions. Each genre has its own register and its own AI-flavour failure mode. The scan workflow stays the same across all five.

CV and resume bullet

The single accomplishment line under a role. Ten to twenty words, action verb up front, a number if there is one. AI rewriters compress real work into a small set of action verbs and lose the specific texture. The fix is to keep the AI-suggested structure but swap the verb and add the actual detail the role required.

Summary paragraph at the top of the resume

The three-to-four-line opening summary is the most templated surface on a resume. Every candidate uses one and most candidates run it through ChatGPT. The result is hundreds of summaries that read alike on a recruiter's screen. Scan the summary on its own and rewrite the lines that flag. A summary that sounds like you raises the read-rate on the rest of the document.

Cover letter

The most personal genre in the packet. Three to five short paragraphs of motivation and connection to the role. AI flavour hurts the most here because the cover letter is the one place a hiring manager expects to hear your actual voice. The opener and the closing paragraph carry most of the risk, and the middle paragraphs about your fit for the role are where the templated language stacks up.

LinkedIn About section

The first paragraph of your LinkedIn profile, read by every recruiter who lands on your page. Most candidates leave this on default or copy the resume summary into it. A distinct About section drafted in your own voice raises the inbound-message rate from recruiters running saved searches in your category.

Portfolio descriptions

For designers, PMs, engineers, and writers running a portfolio site, the short prose under each project entry is read before the work itself. A generic AI-flavoured blurb reduces click-through to the actual project. Scan each description and keep the voice consistent across the portfolio.

Calibration awareness

The narrow action-verb vocabulary that gives AI-rewritten resumes away.

AI-rewritten resumes tend to use the same small set of action verbs across every bullet, regardless of what the role actually involved. Drove, spearheaded, leveraged, optimized, orchestrated, championed, streamlined. The narrow vocabulary is one of the clearest tells on a resume, and the easiest to fix.

Why the verb pool is so small

ChatGPT and similar models were trained on resume samples that themselves used a narrow set of action verbs, and the model defaults to that pool when generating resume bullets. Run a hundred different bullets through any general AI rewriter and the verb distribution collapses onto roughly twenty words. A real writer with actual experience varies the verb across the role: shipped, debugged, hired, escalated, killed a feature, rebuilt a system, lost a deal, recovered an account. The vocabulary stretches because the work is specific.

What the sentence-level scan shows

The scan highlights bullets where the verb-and-structure pattern reads as templated. The fix is not to remove the structure but to swap the verb for one that fits the actual work and add the concrete detail that the AI compressed out. Bullets that score clean after a rewrite read as both polished and specific, which is the combination that lands interview invitations.

Numbers and metrics on their own do not fix it

A common piece of resume advice is to add numbers to every bullet to make it specific. The advice is correct but not sufficient. A bullet that reads "spearheaded a 40 percent improvement in pipeline velocity" still reads templated even with the number, because the verb and the framing are the AI signal. Swap "spearheaded" for the specific verb that describes what you actually did, then keep the number.

Hiring trust

Hiring managers spot ChatGPT-rewritten resumes inside the first ten.

A hiring manager reading the seventh application of the morning notices the third identical opening summary in a row. The reaction is not always conscious, but the interview-invitation rate drops on resumes that read templated. Authentic voice raises the call-back rate even when the underlying experience is identical.

What the hiring manager actually sees

The same opening line across applicants. The same three action verbs under every role. The same rhythm of bullet length. The hiring manager does not always articulate the pattern, but the read is the same: low effort, low investment in the application, possibly low investment in the role itself. The cover letter that follows compounds the impression if it reads templated as well.

The interview-call rate effect

Reply rates and interview invitations drop on resumes that read fully AI-rewritten, even when the candidate experience is strong. The signal the hiring manager takes from a templated resume is that the candidate did not invest the time to write their own materials. Rebuilding the voice in three or four key bullets and the cover letter opener restores the call-back rate without changing the underlying experience.

Senior and executive CVs

For senior and executive roles the review is higher, the document is longer, and the patterns are subtler. Retained search consultants reviewing executive CVs notice voice issues fast, because they read enough CVs in the candidate pool to calibrate against the templated baseline. Executive search firms are explicit about looking for genuine voice on the document and in the first conversation, because the role pays for judgement and judgement shows up in writing.

Honest framing

Not anti-AI. About authentic voice that ATS and humans accept.

AI as an outline tool, a brainstorming partner, a grammar pass, or a way to surface missing accomplishments is fine and most candidates use it that way. The point is that the final document sounds like you, not like the model. The scan tells you which sentences still need that rewrite before submission.

The honest workflow

Write your own draft of the resume and the cover letter first, in your own voice and from your own notes. Run the draft through AI for grammar, structure feedback, or to surface accomplishments you might have left out. Scan the result. The flagged sentences are where the AI pass over-rewrote and pulled the voice towards the templated baseline. Rewrite those sentences in your own register and resubmit.

What this is not

Not a guarantee of being hired. A clean Authenticity Score on the resume does not get you the interview on its own, and the resume is one input among several into the hiring decision. What the scan does is remove the avoidable downside of a templated AI flavour landing your application in a lower review tier before a human reads the actual experience.

The reading-time payoff

A hiring manager spends ninety seconds on a templated application and three minutes on one that reads as written by the candidate. The extra ninety seconds is where the experience gets read, the questions get formed, and the interview invitation happens. The scan-and-rewrite cycle is what shifts you from one read-mode to the other.

FAQ

Job seekers frequently ask.

Will an AI-rewritten resume get flagged by ATS like Workday, Greenhouse, Lever, or iCIMS?
Increasingly yes. Through 2025 the major ATS vendors added AI-content signals to their candidate review layers. Workday surfaces a writing-quality flag on the recruiter view. Greenhouse and Lever both expose third-party AI scores through their integration partners. iCIMS does similar at enterprise tier. A resume that reads as fully ChatGPT-rewritten can land in a lower review tier before a human ever sees it. Pre-scanning your bullets and rewriting the flagged ones is the workflow that keeps you in the top stack.
Do hiring managers actually spot ChatGPT-rewritten resumes?
Senior hiring managers spot the patterns within five to ten resumes. The same opening summary across applicants, the same action-verb vocabulary, the same three-bullet rhythm under every role. Reply rates and interview invitations drop on resumes that read templated, because the hiring manager assumes the candidate did not invest the time to write their own materials. Authentic voice raises the interview-call rate even when the underlying experience is identical.
Why are AI-rewritten resumes so vocabulary-limited?
ChatGPT defaults to a small set of action verbs for resume bullets. Drove, spearheaded, leveraged, optimized, orchestrated, championed, streamlined, and a few more. When you run a resume through any general AI rewriter, you get those verbs across every bullet regardless of what the role actually involved. A real writer varies the verb to match the work: shipped, debugged, hired, escalated, killed, rebuilt. The narrow vocabulary is one of the clearest AI tells on a resume.
Is the cover letter where AI patterns hurt the most?
Yes. The cover letter is the most personal genre in the application packet. It is the one place a hiring manager expects to hear the candidate's actual voice, motivation, and connection to the role. An AI-flavoured cover letter reads as low effort even when the underlying interest is genuine. Scan the cover letter at the sentence level and rewrite the templated lines in your own register. The opener and the closing paragraph carry most of the risk.
What about senior and executive CVs? Are the patterns easier or harder to spot?
Subtler, but executive search firms are actively looking for them. Retained search consultants reviewing senior CVs spend ten to twenty minutes per document and notice voice issues the way a marketing director notices ad copy. Generic strategic-leadership phrasing reads as hollow against the rest of the candidate pool. The fix is the same: scan, rewrite the flagged sentences, keep the genuine voice underneath.
Is TextSight anti-AI for resume writing?
No. AI as an outline tool, a brainstorming partner, or a grammar pass is fine and most candidates use it that way. The point is authentic voice that both ATS and human readers accept. Run your resume through your own draft first, use AI to surface weak phrasings or missing accomplishments, then rewrite the AI-flavoured sentences in your own words before submitting. The scan tells you which sentences need that rewrite.
Which plan should a solo job seeker pick?
Most job seekers run on Free or Starter through the search. Free covers three scans a day with sentence-level highlights, which handles a casual job hunt of one or two applications a week. Starter at $9.99 a month standard, or $7.49 a month on yearly, covers higher-volume application weeks and includes the Chrome extension for scanning inside LinkedIn and Gmail. Pro at $19.99 a month standard, $14.99 on yearly, makes sense if you are also writing cover letters and rewriting the LinkedIn About section across a long search. Business is overkill for individuals.
Does TextSight train on resumes or share them with anyone?
No on both. Scans are private to your workspace and we do not share resumes or any candidate writing. Text submitted for scanning is never used to train the classifier or any other model. This matters because a CV contains personal information, employment history, and sometimes confidential references. The privacy posture is the same on Free, Starter, Pro, and Business.
Related

More for job seekers.

Your search, your voice. Scanned.

Free to try. No card. Most job seekers run a full search on Free or Starter, with the Chrome extension for scanning inside LinkedIn and Gmail.

Start free, no card See pricing
Built for solo job seekers · Sentence-level highlights · Chrome extension on Starter · No training on your writing