Sentence-length variation
How much sentence length changes within this sample.
AI text detector
AI writing? Check it here.
Review a draft, article, or supplied passage before publication. Paste writing or open a document for GPTZero screening and separate local observations. You choose when to send it.
What this tool can tell you
What it can doScreen text with GPTZero after consent and report local writing observations separately, including contrasting evidence.
What it cannot doGuarantee a correct result or prove who wrote something. Detectors can miss AI and flag human writing.
How to use these tools and read your results
Add your writing
A detector scan needs at least 50 words. TXT, DOCX or selectable-text PDF up to 8 MB also work.
Scan paper pages with a camera
Choose how to check
Add your writing to continue.
Your result
See a sample report (opens in a new tab)Your result will appear here
Add your writing, choose how to check, and press the button.
- ✓ A plain-language assessment
- ✓ What it means — and what it doesn’t
- ✓ Suggested next steps you can tick off
- ✓ A report to keep (your text is excluded)
Compare two drafts
Revision evidence
Compare how two drafts changed
Both drafts remain in browser memory. This describes changes; it does not identify a writer, editor, model, or tool.
Browser memory only · no upload · no retention · source text excluded from export
Research results and measurement glossary
Bounded research snapshot
What the paired pattern did in recent tests
These are aggregate pattern counts from fixed, bounded studies. They are not accuracy, probability, sensitivity, specificity, or proof of authorship. The rule was not changed after these results were reviewed.
Document typeTransformed synthetic samplesHuman reference samples
Essay5 / 6 showed the pair0 / 30 across 10 source groups showed the pair
Article5 / 6 showed the pair0 / 17 across 8 source groups showed the pair
Review6 / 6 showed the pair4 / 30; author grouping unavailable showed the pair
Important: Human reviews also produced the pattern combination. Editing, genre conventions, translation, templates, and individual style can create similar measurements. Review results therefore need especially cautious interpretation and provenance checks.
Privacy and methodology
What the measurements mean
Measurements describe visible structure and wording only. The paired-pattern review is document-aware, abstains where evidence is too limited, and must be considered alongside counterexamples and provenance.
Vocabulary distribution
How broadly words are distributed across moving passages.
Repeated phrases
Repeated three-word sequences; topics, quotations, and templates can create them.
Transitions and hedges
Linking and qualifying words that may reflect genre, editing, or style.
Template-shaped phrases
A small fixed list of common formulaic phrases, never a universal signature.
Citation-shaped patterns
Reference-like forms, not validation that a source is accurate.
Style shift
A change in average sentence length between halves of the document.
Evidence state
A review state—not a classification of who or what wrote the text.
Review desk: Why do AI detectors disagree? →