AI Content Detector

No. The AI Content Detector returns an advisory likelihood estimate, not proof. Use it as one review signal alongside human judgment, drafting history, citations, source context, and editorial review.

It reviews visible writing signals: repeated phrasing, sentence rhythm, generic wording, lack of concrete specificity, formulaic structure, and unusually consistent style. The result explains these signals instead of returning only a label.

Segment notes show where individual paragraphs look more human-written, mixed, or AI-assisted. They help you inspect the evidence locally instead of treating the whole document as one flat score.

Yes. Human writing, second-language writing, translated text, edited AI drafts, templates, and polished business or academic writing can all produce mixed or misleading signals.

Each check shows an AI probability with a matching human probability, plus a confidence level of low, medium, or high. Low confidence means the signals are mixed and the estimate should be treated with extra caution.

Balanced is the default review for everyday checks. Strict scrutinizes the text more closely and flags borderline patterns more readily, which is useful for a closer second opinion but can raise false positives.

Yes. It detects the dominant language and returns a language-aware summary and segment notes for English, Turkish, and mixed-language drafts, instead of assuming every text is English.

No. The result is an advisory estimate, not evidence. It should never be the sole basis for academic, hiring, or disciplinary decisions — weigh it alongside drafting history, sources, conversation, and human review.

Signed-in free accounts get monthly AI trial actions. Pro users can use the larger monthly AI limits on their plan.