ScribeLens checks text for the writing patterns typical of ChatGPT and similar assistants, then shows you the specific sentences and signals behind the result rather than a single opaque percentage.
Large language models produce text with characteristic tendencies: consistent sentence rhythm, an even level of formality, predictable transitions, and a preference for balanced, hedged phrasing.
None of these are unique to AI — plenty of careful human writing shares them, which is exactly why evidence matters more than a verdict.
Tools that claim to identify exactly which model produced a text are overstating what can be measured. Modern assistants produce broadly similar surface patterns, and output changes as models are updated.
What can be assessed is whether text shows patterns common to AI-assisted drafting in general. That is what ScribeLens reports, and we would rather describe it accurately than market a capability that does not exist.
Start with the sentence-level view. Concentrated flags in one section suggest a different story than uniform flags throughout, and reason keywords tell you which pattern triggered each signal.
Whatever the score, treat it as a prompt to look closer rather than a conclusion.
It detects patterns common to AI-assisted writing, including text from ChatGPT. Claiming to identify one specific model would overstate what pattern analysis can reliably measure.
Editing reduces the patterns detectors rely on, so heavily rewritten AI text may read as human. This is a limitation of the whole category, and why a low score is not a clearance.
Yes. Guest checks require no signup, and a free account provides 35,000 tokens per month, where one token covers one word.
Longer passages produce more reliable signals. Very short excerpts do not give pattern analysis enough to work with.
Paste a draft and see the sentence-level evidence behind the result. Free, no signup required.