A hidden text watermark is not a visible character, footer or file tag. It is a statistical pattern created by the model's token choices, so copying the text into a new document may leave that pattern intact.
IA Checker's Watermark Remover performs a substantial rewrite of pasted text or an uploaded document. It changes phrasing, sentence rhythm and token distribution while preserving meaning, facts, names and quotations. Those are the same structural changes that can disrupt a statistical watermark, rather than merely hiding a visible label.
After rewriting, run the result through the AI Detector to review its sentence-level authenticity signals. IA Checker cannot yet issue a provider-verified watermark-removal certificate because Anthropic has not published an independent Claude watermark verifier. The rewrite workflow is available now; provider-specific verification can be added when that mechanism becomes public.
