NewOctober 5, 2026: OpenAI announces text watermarking for ChatGPT models. Learn more →
Clean AI WatermarkGemini has carried Google DeepMind's SynthID-Text watermark since 2024. Paste the text below to break the pattern and keep what you wrote.
Paste on the left and hit Remove watermark. Takes about ten seconds.Paste above and hit Remove watermark. Takes about ten seconds.
SynthID-Text is the original word-pattern watermark. Google DeepMind published it in Nature in October 2024 after running it live across close to twenty million Gemini responses. A secret key scores candidate words as Gemini writes and nudges it toward the higher-scoring ones. The text reads normally, but a detector with the key sees the tilt across a few hundred words.
We swap just enough of the steered words for equivalents to break the pattern, and remove zero-width characters, soft hyphens and markdown leftovers in the same pass. Sentence structure, meaning, length and tone stay the same.
No. It is a statistical pattern in word choice. Nothing in the text looks different, and no hidden character marks it.
Heavy rewriting or translation does, but it also changes your text. Clean AI Watermark swaps only as many words as the pattern needs, so the text stays yours.
Detection needs the key. Google holds it, so the provider and the partners it grants access to can run the detector; the public cannot.