Paste a LinkedIn post, an email, or a page. The checker marks the patterns that make writing read like AI and says how to fix each one. It runs in your browser and keeps nothing.
| Em dashes | a style marker in Wikipedia's guide to AI writing |
|---|---|
| "Not just X, it's Y" | the negate-then-assert sentence |
| Hype words | unlock, seamless, leverage, game-changer, and the rest |
| Signposting | "In conclusion", "Here's the thing", "Let that sink in" |
| Hashtag walls | four or more tags in a row |
| Flat rhythm | five or more sentences of nearly the same length |
The rules come from the open-source humanizer projects by blader and jcarterjohnson (MIT) and Wikipedia's “Signs of AI writing” guide. We run the same checks on everything we publish.
No. It finds patterns that readers associate with machine-written text, and people write them too. A clean result means none of these patterns showed up, not that a person wrote it.
No. The checker runs in your browser. Nothing you paste is sent to us or to any model.
Wikipedia's guide to spotting AI writing lists overuse of em dashes among the style markers of machine-written text. A comma or a period does the same job and removes the doubt.