Detect AI-generated text with transparent multi-signal analysis. Scans for AI buzzwords, sentence variety, em-dash patterns, and vocabulary diversity to produce an explainable probability score. Works offline with local rules, or connect an AI provider in settings for hybrid analysis with enhanced accuracy.
No tool can guarantee perfect accuracy. This detector provides a probability score based on linguistic patterns, not a definitive verdict. Use it as a starting point for review, not as conclusive proof of AI authorship.
No. The tool works fully offline using local rule-based analysis. Optionally connect an AI provider in settings to enable hybrid mode, combining local signals with AI-powered assessment for potentially higher accuracy.
It checks five key signals: AI buzzword density, sentence length variety, em-dash usage patterns, transition word frequency, and overall vocabulary diversity. Each signal displays its score and example evidence from your text.
The detector identifies common AI writing patterns rather than specific models. It flags text with characteristics frequently found in AI-generated content, such as overused phrases and uniform sentence structures, regardless of the source model.