How the AI text checker works
We split your text into sections and run each through an open-source language model trained to tell human writing from the output of modern chatbots (ChatGPT, Claude, Gemini, Llama and others). You get an overall AI score plus a colour-coded breakdown, so you can see exactly which parts look AI-written. We also point out phrases chatbots overuse — like “delve”, “tapestry” and “it’s important to note”.
Use it fairly
AI text detectors are much less certain than image labels. Non-native English, very formal writing, lists and templates can look “AI-like”, and lightly edited AI text can look human. Never use a detector score alone to accuse someone of cheating — use it as a reason to look closer and talk to them.
- Paste at least 150 words for a meaningful result (40 minimum).
- Check sections separately if only part of the text seems suspicious.
- Look at the writing process: drafts, version history and sources tell you more than any score.
Frequently asked questions
Can it detect ChatGPT writing?
Yes — the detector was trained on text from ChatGPT and other popular chatbots. Very short or heavily edited text is harder to judge, so the result includes a confidence level.
How much text can I check?
Up to 15,000 characters (about 2,500 words) per scan, and as many scans as you like.
Is my text saved?
No. Text is analyzed and discarded. If you choose to share your result, only a short excerpt is stored.
Is IsItMadeByAI really free?
Yes. Every scan is free, there is no sign-up, no credit card and no watermark on your results. The site runs on open-source detection models, which is how we keep it free.