Paste any text below and get a detailed linguistic analysis — probability scores, style metrics, and patterns that reveal whether a piece of writing shows the characteristics of automated composition.
Start AnalysisNo signup required • Results in seconds • Up to 8,000 characters per check
Examining linguistic patterns...
--
Scores are statistical estimates, not definitive proof. Writing style, editing quality, and text length all affect results.
Before you turn in a paper, scan your draft. Identify sentences that read unusually flat or formulaic, and polish them so your own voice comes through clearly.
When grading feels ambiguous, a quick scan gives you another angle. Look for unusually consistent sentence length or vocabulary patterns to guide conversations with students.
Review incoming articles and commissioned pieces. Identify sections that may need rewriting to feel more conversational and less mechanically produced.
Working with AI-assisted drafts? Run your own pieces through the analyzer before publishing. It helps keep brand voice authentic and editorial quality high.
The analyzer is designed as a second pair of eyes — not a substitute for careful reading. It highlights patterns you can then investigate yourself.
After you submit text and tap analyze, the system begins parsing the submission at multiple levels simultaneously. It measures sentence length variation, counts repeated phrases, examines punctuation rhythm, and traces how ideas connect from one paragraph to the next. Human writing tends to show more of what statisticians call burstiness — long sentences next to short ones, unusual word choices, minor inconsistencies that reveal a person behind the keyboard.
The analyzer then compares what it found against reference data built from many sources — published articles, submitted essays, research papers, and outputs observed from various language models. This reference set is updated as new models are released, so the tool keeps pace with developments in automated writing rather than relying on a frozen snapshot of older capabilities.
The final step turns numbers into something you can act on. You see two percentages expressing the likelihood of automated versus human composition, several plain-language metrics describing sentence length and vocabulary diversity, a short written summary of findings, and a list of specific patterns flagged during the scan. The goal is to give you a readable report rather than a mystery score.
Analysis runs on demand, and nothing you paste is retained after you leave the page.
This is a free browser-based tool for examining writing style. It does not search the internet or compare against an existing database — instead, it looks at the characteristics of the writing itself: sentence structure, vocabulary choice, repetition patterns, and the flow of ideas. The goal is to give readers a clearer picture of how a piece of text was composed, without making absolute claims about its origin.
A standard plagiarism checker looks for verbatim copies elsewhere online. That approach cannot catch content that has been paraphrased or written from a prompt. This analyzer works differently — it looks at the statistical signature of the writing and flags things that commonly appear in text produced by language models. It works best on English text of several paragraphs or more, where there is enough material to establish a writing profile.
Using the tool is straightforward: paste your text and tap the analysis button. You will receive results within a few seconds. No account is required, and no submitted material is saved after the page closes. Remember that results are informational and should not be used alone as the basis for academic decisions or professional judgment.
No account is needed. You can begin analyzing text immediately after loading the page. There is no email sign-up, no subscription, and no premium tier that blocks core features.
Your submission is processed immediately and discarded when your session ends. The system does not maintain a library of what users paste, does not share submitted text with third parties, and does not log individual analyses.
The tool provides statistical estimates based on writing patterns, not absolute judgments. Some writing — highly edited text, short paragraphs, or very formal prose — can be harder to classify. Think of the output as information to consider, not a definitive answer.
You can paste up to 8,000 characters per check — roughly 1,200 to 1,500 words. Longer pieces give the analyzer more material to work with and tend to produce more stable observations. If your text is longer, break it into pieces and scan each section separately.
Most analyses complete in a few seconds. Shorter submissions return even faster. If the tool is under heavier-than-usual traffic, processing may take slightly longer but should still feel responsive.
The analyzer is calibrated for English writing. Other languages will still produce a report, but results may not be as reliable. For best results, use it with English text.
Many educators and editors use it as an additional signal alongside their own reading. That said, results should never be used as the sole basis for an academic or professional decision. Always combine them with your own review of the text.
The tool allows a reasonable number of daily scans per visitor to keep performance stable for everyone. If you have a larger workflow — for a classroom, publication, or team — please get in touch to discuss enterprise options.