ChatGPT watermark remover
Paste text from ChatGPT and it names every hidden character in there, then hands back a clean copy.
Nothing you type here leaves your device. There is no upload and no request.
Files carry their own
This handles text you paste in. Images and documents keep their hidden data somewhere else entirely — in the file’s metadata rather than in its characters — including the C2PA provenance record that AI image tools sign into what they generate. The metadata tools here strip that, and the pixels come through untouched.
What the ChatGPT watermark remover finds in your text
Text that comes out of ChatGPT carries characters you cannot see. The best known is U+202F, a narrow no-break space, which turns up where an ordinary space belongs — people began noticing it in GPT-o3 output in April 2025 and the name stuck. Alongside it come no-break spaces, the typographic spaces a model picks up from well-set training text, soft hyphens, and occasionally zero-width characters. This walks your text one code point at a time, sorts what it finds into eight categories, and reports each one by name and count before handing back the cleaned version. Anything invisible with no width is deleted; the unusual spaces are replaced with an ordinary space, so the words on either side stay apart.
What it cannot do
- It works on text you paste in. Images and PDFs carry their own metadata, which the metadata tools here handle.
- Formatting that relies on these characters goes with them. A no-break space holding a number next to its unit becomes an ordinary space, so that line can break again.
- It reports what a character is, not how it got there. A zero-width space in pasted text is far more often a word processor artefact than anything deliberate.
Questions people actually ask
- What is the ChatGPT watermark?
- In practice people use the phrase for the invisible characters that show up in ChatGPT output — above all U+202F, a narrow no-break space appearing where a plain space should be. It became widely noticed in GPT-o3 output in April 2025. Those characters are genuinely present in the text, and this removes all of them.
- How do I know it found anything?
- It shows you. Every category is listed with a count — narrow no-break space times seven, soft hyphen times two — before you copy anything, so you can see exactly what was sitting in your text.
- Why do these characters matter?
- Because they break things quietly. A no-break space fails a string comparison, so a search box will not match text you just pasted. A CSV column splits in the wrong place. A compiler rejects a line that looks perfectly fine. Cleaning them is worth doing whatever put them there.
