Anthropic just

Anthropic just tagged every file Claude makes so it can never pass as human work. I'm not even exaggerating. And someone already built the fix. So here's the deal. Anthropic just confirmed every file Claude generates chats, co-work, Claude code, all of it gets a hidden signature baked in, in the corner. Not a logo in the corner. We're talking invisible unicode and signed metadata. Your PNGs, PDFs, SVGs, HTML, basically anything Claude spits out. Then, 24 hours later, a dev drops a repo called Watermark Remover and it strips the hidden metadata out of your files. No more "Made by AI" tag riding along. Here's the part that makes this insane. It doesn't just work on Claude. Same tool wipes the fingerprint from ChatGPT and Gemini outputs to one tool, every major AI. Pretty freaking cool. Now, worth knowing the file metadata is a clean strip. Text is trickier since that watermark lives in word choices, not a tag. Still, for images and docs, gone. Gone in one pass. Check my caption below for the GitHub repo link. 👇 Anthropic has confirmed that files generated through Claude can include provenance information that helps identify them as AI-generated. We're not talking about a visible logo. The information can be attached to files like: • PNGs • PDFs • SVGs • HTML Then, less than 24 hours later, a developer released Watermark Remover, a GitHub tool that can strip certain metadata and file properties from supported files. It also works with files generated by ChatGPT and Gemini.