Paste a text you are not sure about and see which sentences read like output from an AI chat. The demo below checks a section of up to 3,000 characters without signing in; with a free account you see all seven detectors side by side.
Six more detectors
ChatGPT writes evenly: sentences of similar length, paragraphs with the same structure and phrases like “in today's world”, “it is important to note” or “in conclusion”. Detectors look at exactly these properties, along with how predictable the word choice is.
Other large models write the same way, so the check recognises machine writing in general — text from Claude or Gemini comes out much like text from ChatGPT.
Which model wrote the text. The differences between models are smaller than the differences caused by the prompt or by later edits, and an independent detector has no key to their watermarks. OpenAI has not publicly launched text watermarking for ChatGPT.
Nor can it reliably recognise text that someone thoroughly rewrote, translated or mixed with their own after generating it. With text like that the detectors disagree and the result is borderline.
Few texts come entirely from a chat. More often it is a person's own text with a few generated paragraphs, or chat output with edits of their own. That is why you should read the marked sentences, not just the number — they show where the style changes.
When you check your own text, Style deviation helps too if you have an account: it compares the text with your earlier texts and highlights the sections that differ from them most.
A high score is not proof that ChatGPT wrote the text, and a low one is not proof of the opposite. No AI text detector is one hundred per cent accurate.
When you check someone else's text, take the result as a basis for questions: why the author chose this source, what the term in the second paragraph means, how they reached the conclusion. When you check your own, it shows you the passages worth rewriting in your own words.
Yes. The check measures the properties of machine writing in general, not the signature of a particular model, so text from Claude or Gemini comes out much like text from ChatGPT.
OpenAI has not publicly launched text watermarking for ChatGPT. Invisible characters can still get into a text when it is copied, though, and the check with an account looks for them separately.
At least 100 characters; a continuous paragraph or a whole chapter is more reliable. With just a few sentences the result is indicative only.
Proofreading changes a text only a little, so it usually comes out just slightly higher. But if ChatGPT rewrote whole sentences, the text can come out high even if the work is your own.
Texts up to 300,000 characters, seven detectors at once, colour-marked sentences and a history of your checks. The account is free and needs no card.