AI Text Detector — Free ChatGPT & GPT Checker
Paste at least 300 words (up to 150,000 characters) or upload a document to get an AI-likelihood score, highlighted sentences and a plain-language verdict.
Limitations: Results are probabilities, not proof. Like every AI detector, this one can be wrong, including flagging human writing as AI. Don’t use a score on its own to accuse anyone. How results can be wrong →
Paste your text to scan
Paste Text to Scan
How the AI Text Detector Works
Your text is sent to a third-party AI-detection API, which looks for statistical patterns of machine-generated language — word choice, sentence rhythm and predictability that tend to differ from human writing. Results come back as an overall confidence score plus sentence-level flags, so you can see exactly which parts of a passage drove the verdict instead of trusting a single number.
Sentence-Level AI Detection
The result highlights the sentences that drove the score, so you can see which parts to look at more closely. A highlight is a reason to look, not proof.
English Only
The site supports English only. Other languages aren’t supported, and results on non-English text shouldn’t be relied on.
Frequently asked questions
Results are probabilities, not proof. Like every AI detector, it can be wrong in both directions: it can miss AI-generated text, and it can flag human writing as AI (a false positive). False positives are more likely on short, formulaic or heavily edited text and on writing by non-native English speakers. Treat a score as one signal, never as a verdict.
No. The site supports English only. Results on other languages aren't supported and shouldn't be relied on.
Paste or upload up to 150,000 characters per scan. A scan needs at least 300 words, because short passages don't give the detector enough signal. Uploads can be .txt, .md, .docx or .pdf files up to 8 MB.
Your text is sent to a third-party AI-detection API to produce the result. If you're signed in, your scan history keeps the first 200 characters of the text and the result; signed-out visitors have no server-side history. The privacy policy lists every provider.
It can flag text that shows statistical patterns common to large language models such as ChatGPT. It can't confirm which tool, if any, wrote a passage, and it can both miss AI-written text and flag human-written text.
Yes, within the free tier: 3 free scans in total, shared across all five tools (not per day or per month). No signup needed to start. Paid plans remove the scan limit for AI detection.
They are the sentences that contributed to the overall AI-likelihood score, so you can see what drove the result instead of trusting a single number. A highlighted sentence is not proof that it was AI-written.
Less so. Text that mixes human and AI writing, or has been edited, paraphrased or translated, produces less reliable scores, and fully human writing can still be flagged. Use the score as a reason to look closer, not as a conclusion.
Not on its own. Detector results can be wrong, including on genuine student work. If a result raises a concern, look at drafts and revision history and talk with the writer before drawing any conclusion.