OpenAI is now embedding an invisible watermark, called textGrain, in text from ChatGPT and Codex. It's on by default across every plan in the EU, and any API customer, anywhere, can switch it on. The ChatGPT Watermark Detector checks a piece of text for the mark and provides a clear reading, so you don't have to guess whether something is clean before you publish it or trust it.
Words limit: 0 / 500
Everything you need to check, verify, and monitor ChatGPT's textGrain watermark, whether it's a single page or your entire publishing workflow.
One document at a time. Paste in any piece of text, ChatGPT's or otherwise, and get a transparent read on where it stands. No account setup needed to try it, and no limit on how many times you check the copy before you're confident in the result.
No unclear "it might be AI" language. Every check comes back with a detection score attached, so you're confident instead of eyeballing it and hoping.
Route content through the Watermark Detector API before it hits your CMS, your inbox, or your archive. Built for teams drafting in ChatGPT or Codex who need to check at scale, whether that's verifying your own rewritten content stays clean or screening incoming text at volume, with fast turnaround inside a real production pipeline.
What is ChatGPT textGrain, and can you detect it?
textGrain is not a logo or mark at the corner of an image. As ChatGPT writes, it scores every possible next word, and textGrain uses a secret key to tilt those scores in a consistent pattern. Nothing is added to the page, no hidden characters and no metadata, so it's invisible and travels with the text through copy and paste. OpenAI designed it to hold up through copy and paste and small edits. It's on by default across every ChatGPT and Codex plan in the EU, and any API customer, anywhere, can switch it on.
OpenAI has also published where the signal breaks: in its own tests, swapping 10% of the words for synonyms dropped detection from about 92% to 66%, and short passages, math answers, translated text, and heavy edits all weaken it further. The Watermark Detector is built around that word-choice signal, not the surface-level style guesses older AI detectors rely on. It reports a clean result: no textGrain found, not proof a human wrote it.
Signal checked
Hidden characters
Result format
Volume
Update cadence
Still have more questions?
Learn more in our support page.
Paste the text into the ChatGPT Watermark Detector. It checks for textGrain's word-choice signal and returns a detection score, so you know whether the mark is present before you publish the piece or trust it.
Yes, when textGrain is on. OpenAI embeds an invisible watermark called textGrain in text from ChatGPT and Codex. It's on by default across every plan in the EU, and any API customer, anywhere, can switch it on. The mark is a pattern in word choice, not a visible logo, hidden characters, or metadata.
Yes. This tool is built around textGrain. A clean result means textGrain wasn't found, not that no other model was involved.
No. A missing mark doesn't automatically mean a piece is human-written. Short passages, synonym swaps, translation, and heavy edits can all weaken or break the signal. The detector reports a clean result as no textGrain found, not as proof of human authorship.
Yes. That's one of the main uses. Run a draft through the detector before it hits your CMS or publish button so you know whether textGrain is still in the text.
A general AI detector guesses from style and phrasing. The Watermark Detector checks for textGrain's actual word-choice pattern and returns a score, rather than a vague "might be AI" label.
Yes. The ChatGPT Watermark Detector checks text for OpenAI's invisible textGrain mark and returns a detection score on every check.
You can't prompt ChatGPT to leave the mark out. textGrain is embedded as the text is generated. OpenAI's own tests show a light synonym swap still leaves most of the signal. To strip it, you need a structural rewrite. StealthGPT's ChatGPT Watermark Remover is built for that.
One person checking one document, a team verifying content before it goes out, or screening submissions at scale, the check works the same way. Enterprise options are available for teams that need it running quietly in the background.
