Just no. The EU literally just passed regulations requiring AI-generated content to be clearly labeled, and youre out here trying to make the internet even more sloppy and untrustworthy than it already is, cool project....
Good for the EU. I am not subject to their authoritarian decrees, thank God.
This is a privacy oriented web utility for removing claude's watermarks from its text outputs and image outputs.
Image files are processed locally in the browser. It removes the C2PA signals from them directly.
Text inputs use an open model (Nemotron) on OpenRouter with Zero Data Retention enabled, no logging of inputs or outputs is done on this site or by OpenRouter or their providers.
It is 100% free to use and is ad-network free.
How do you detect the watermarks and how do you remove them. Code please.
Right now we use a special algorithm to rephrase the submitted text without changing or losing meaning. Once Anthropic documents how to detect the watermark itself work to remove it more surgically can commence.