Free TextGrain Remover
Rewrite and clean AI-generated text while keeping its meaning. No sign-up, no email.
Updated October 7, 2026
Clean only runs in your browser. Rewrite sends the text to an AI model.
Rewrites text using a separate AI model and removes hidden formatting artifacts. OpenAI does not offer public textGrain verification, so removal cannot be guaranteed.
What is OpenAI textGrain?
textGrain is OpenAI's watermark for AI-written text, announced on October 5, 2026. It doesn't add anything you can see or select. When several words would fit a sentence equally well, a secret key nudges which one the model picks. Over a long enough passage those choices form a pattern that a detector holding the same key can recognize.
OpenAI is adding it to text from eligible ChatGPT and Codex users on all plans in the EU over the coming weeks, to meet Article 50 of the EU AI Act and the EU Code of Practice on transparency of AI-generated content. Outside the EU it's off in ChatGPT at launch, and API customers can opt in for select models.
- AnnouncedOctober 5, 2026
- WhereChatGPT and Codex in the EU; opt-in in the API
- How it worksStatistical pattern in word choice, no hidden characters
- DetectorApproved researchers and organizations only
How this TextGrain remover works
Most watermark removers only strip hidden characters. textGrain isn't made of characters, so this tool changes the wording itself, then shows you what changed.
- 01
Clean
Strips zero-width characters, direction marks, hidden tag characters and odd spacing in your browser. Free, instant and private.
- 02
Rewrite
A separate AI model rebuilds the wording sentence by sentence while keeping every fact, number, name, link and quote.
- 03
Check
You see how much wording changed, which hidden characters were removed, and whether any concrete detail went missing.
Can you remove a textGrain watermark?
Editing weakens it, by OpenAI's own numbers. On 400-token passages, swapping 10% of the words for synonyms cut detection from about 92% to 66%, and swapping 25% cut it to about 17%. OpenAI also says substantial paraphrasing or translation can make the watermark undetectable.
What nobody outside OpenAI's approved list can do is check the result, because the detector isn't public. That's why this tool reports how much of the wording changed instead of a “watermark removed” score. Any site showing a removal percentage for textGrain is guessing.
92%
unedited
66%
10% of words swapped
17%
25% of words swapped
Share of 400-token passages OpenAI's detector flagged, as reported by OpenAI.
Hidden characters vs. textGrain: what’s the difference?
People call both of these “AI watermarks”, but they work in completely different ways, so they need different fixes. This tool handles both.
Copy-paste leftovers
Hidden characters
- What it is
- Invisible symbols, like zero-width spaces, that can ride along when you copy text out of a chat app.
- Can you see it?
- No, but any software that checks text can find them in seconds.
- How this tool handles it
- Clean only deletes them in your browser. Done, and easy to confirm.
OpenAI's watermark
textGrain
- What it is
- A hidden pattern in which words ChatGPT picked. Nothing extra is added to the text.
- Can you see it?
- No. Only OpenAI's detector can check for it, and that detector isn't public.
- How this tool handles it
- Rewrite & clean rewords the text. More rewording weakens the pattern, but nobody outside OpenAI can confirm it's gone.
TextGrain questions
Yes, in the EU. On October 5, 2026 OpenAI said it will add its textGrain watermark to text from eligible ChatGPT and Codex users across all plans in the EU over the following weeks. Outside the EU, ChatGPT text is not watermarked at launch. API customers anywhere can opt in for select models; it is off by default.
No. textGrain is a statistical pattern in which words the model picks. Where several words would fit equally well, a secret key nudges the choice, and a detector holding the same key can spot the pattern across a long enough passage. It adds no hidden characters, invisible spaces or unusual punctuation, so tools that only strip hidden characters do not touch it.
No, and no tool can promise that today. OpenAI's detector is not publicly available, so nobody outside approved organizations can check whether a passage still carries the signal. This tool rewrites the wording with a separate AI model, which is the kind of change OpenAI reports weakens detection, and it tells you exactly how much wording changed.
Not yet. OpenAI is taking applications for detector access, initially limited to approved researchers and expert organizations. Any site that claims to detect or verify textGrain removal for the public is not using OpenAI's detector.
OpenAI's own testing on 400-token passages found detection fell from about 92% to 66% when 10% of the words were swapped for synonyms, and to about 17% at 25%. It also says substantial paraphrasing or translation can make the watermark undetectable. Detection is strongest on long, unedited text: about 80% of 200-token passages and 95% of 400-token passages at a 1% false-positive target.
It removes characters you can't see that often ride along when text is copied from chat tools: zero-width characters, direction marks, hidden tag characters, soft hyphens and stray variation selectors. It also turns unusual spaces into normal ones and tidies extra whitespace. It runs entirely in your browser and does not change any words.
WildSEO does not store the text you paste. Clean only never leaves your browser. For a rewrite, the text is sent through OpenRouter to the AI model and the result comes straight back to you. To prevent abuse we log a salted hash of your IP address, the time and the length of the text, never the text itself.
Yes. There is no sign-up and no email wall. Each rewrite takes up to 8,000 characters, about 1,300 words, and there is a limit on rewrites per hour to keep it free for everyone. Longer pieces can be rewritten section by section.
Use it on your own drafts. If your school, employer, publisher or local rules require you to disclose AI assistance, rewriting the text does not change that obligation.
Related reading
What is textGrain? Full definition, how detection works, and how to handle it in a publishing workflow.
