Claude Watermarks and How To Remove Them

Anthropic has signed the EU AI Act's Article 50(2) Code of Practice on Transparency of AI-Generated Content. As a provider of both generative AI models and generative AI systems, Anthropic is now putting transparency commitments into practice through machine-readable marks in Claude-generated content. This article summarizes how marking works, where it applies, and what its limitations are.

Anthropic's Marking Commitments

Under the EU AI Act's Code of Practice, Anthropic has made four key commitments for Claude:

• New models will mark AI-generated content from day one. Claude models launched in the EU on or after August 2, 2026 support machine-readable marking at launch. Generated text carries embedded watermarks, and generated files include digitally signed provenance metadata where supported.
• Marking works everywhere you use Claude. Marks apply to output from supported Claude models across Claude Platform (API), Claude, Claude Code, Claude Cowork, and Claude Tag — wherever Claude is offered, worldwide. Some platforms or features may not support certain marking types.
• Anthropic will help users detect Claude's marks. The company will support users and third parties in detecting Claude's marks, as the Code requires, with details shared in forthcoming documentation.
• Existing models are in progress. The law includes a transition period for Anthropic models launched before August 2, 2026, and marking support is being added to those models as well.

AI transparency commitments are reshaping how generated content is labeled and tracked.

What's Covered

Marking applies across several dimensions:

• Models: Claude models launched on or after August 2, 2026 support marking at launch. Older models are being updated.
• Products: Claude markings cover output from supported models everywhere Claude is used — Claude Platform (API), Claude, Claude Code, Claude Cowork, and Claude Tag. Embedded watermarks apply to all generated text. Provenance metadata applies where Claude supports processing files.
• Cloud partners: Embedded watermarks apply when supported Claude models are accessed through AWS, Google Cloud, or Microsoft Foundry. Signed provenance metadata may not be supported on every platform, depending on the features each platform offers.
• Regions: Marking applies to output from supported models wherever Claude is offered, worldwide.

How Claude Marks Content

Claude uses two complementary techniques to mark content it generates and processes: (1) watermarks embedded in text, and (2) signed provenance metadata attached to files.

Embedded watermarks are woven into generated text invisibly, traveling with copies and some edits.

1. Embedded Watermarks in Text

When a supported Claude model generates text, it weaves an imperceptible watermark directly into the text itself. You won't see it, and it doesn't change the meaning, quality, or readability of Claude's response. Because the watermark is part of the text, it travels with the text when copied and pasted elsewhere, and may persist through some editing. Watermarking is applied at the model level, meaning it is present no matter which Claude product or surface the text comes from.

2. Signed Provenance Metadata

When Claude generates a supported file type — such as .svg, .png, or .jpg — it attaches signed provenance metadata. This metadata follows the Coalition for Content Provenance and Authenticity (C2PA) open standard, which is used across the industry to record information about content provenance. If a signed metadata label is present, it signals that a file was processed by Claude and lets you detect whether the file has been tampered with.

Detecting Claude's Marks

Anthropic is working to enable users and other third parties to detect Claude's embedded watermarks and provenance metadata. Detection checks whether a piece of text or a file carries a supported Claude mark. If a supported mark is found, it indicates that the content may have been processed by Claude. Detailed detection mechanisms will be shared in forthcoming technical documentation.

Detection signals are useful but not conclusive — marks can be stripped or may never have been present.

Limitations

Machine-readable marks provide important signals about content, but they have meaningful limitations across all content types.

A detected mark provides a signal, but is not fully conclusive. Detecting a Claude mark tells you that the content may have been processed by Claude. It does not, on its own, confirm the full provenance of the content. For example:

• Claude may not be the original author. People often use Claude to proofread, translate, summarize, or convert files. The output can carry a Claude mark even if the underlying ideas, text, or data originated from another source.
• The content may have changed after Claude processed it. Marked content may be modified, excerpted, or combined with other material after Claude processed it.

Lack of a detected mark doesn't mean the content wasn't AI-generated or processed. Content generated by Claude may not carry a detectable mark if:

• It was generated by a model released before marking was supported;
• The text has been heavily edited, paraphrased, translated, or mixed into other writing;
• The passage is very short, leaving too little text for a reliable signal;
• A file's metadata was stripped through format conversion, re-saving, screenshots, or other means;
• It was produced through a platform, feature, or file type where a particular marking type wasn't supported.

Developers deploying Claude must assess their own transparency obligations under the EU AI Act.

If You Build With Claude

If you deploy Claude in your own product, you should independently assess what Article 50 requires of your products and services. Consistent with Anthropic's commitments under the EU Code, the goal is to support developers in meeting their own transparency obligations. Technical guidance on marking and detection will be shared as it becomes available.

Try the Watermark Removal Tool

Curious about how watermark detection works in practice? Paste a piece of text below to test our watermark removal tool. This is a demand-test tool — we're gauging interest in watermark detection and removal capabilities before building a full-featured product.