AI detector
Score your text before and after a rewrite. Sign in to scan. The free plan includes 1,500 detector words a month.
Detector scores are estimates from third-party classifiers and change as those tools update. No text can be guaranteed to pass any detector. Use this to improve your writing, and follow your institution's or employer's rules on AI use.
How AI detectors work
Most detectors estimate two things: how predictable each word is to a language model (often called perplexity) and how much that predictability varies across a text (burstiness). Model-written prose is uniformly predictable; human prose is uneven: a short sentence after a long one, an unexpected word, a concrete example. Newer detectors add trained classifiers on top. All of them output a probability, and all of them are updated without notice, which is why the same text can score differently from month to month.
Why false positives happen
Anything that makes writing more predictable raises the estimate: short length, formulaic genres (lab reports, legal text, listicles), heavy use of grammar tools, and writing by non-native English speakers (a documented pattern across vendors). That is why responsible vendors tell institutions not to act on a score alone, and why we show the score as feedback for your editing rather than as a verdict.
Detector guides
- GPTZeroWhat GPTZero measures (perplexity, burstiness, classifier signals), how to read its document and sentence scores, and why false positives happen.
- TurnitinHow Turnitin's AI indicator differs from its similarity report, what the percentage means, its published limits, and why institutions treat it as a signal, not proof.
Questions
Is the AI detector free?
Yes, with an account: the free plan includes 1,500 detector words a month (500 per scan). Paid plans add larger allowances and a multi-detector view.
How accurate are AI detectors?
They estimate predictability and uniformity in text. Vendors report high accuracy on benchmark sets, but real-world results vary by length, genre, editing and the writer's background, and every vendor documents false positives. Treat a score as a signal.
Why was my own writing flagged?
Short, formulaic or very polished text (and text by non-native English writers) is more predictable to a model, which detectors read as machine-like. Adding specifics and varying sentence rhythm usually lowers the score.
Which detector does AntiGPT use?
An open classifier we host, labelled as an estimate; paid plans can add a second opinion from a commercial detector. Scores are estimates and are never presented as a guarantee.