Guides
Claude Watermark Guides
Claude watermark guides on this page answer the questions that decide whether any of this affects you at all. Every claim here is sourced from Anthropic's own documentation or from primary reporting, and where the honest answer is "nobody knows yet," that is what it says.
Runs in your browser. No signup, no upload, nothing stored.
How it works
- Claude Watermark: How Does It Work?Text: biased token sampling keyed to a secret. Files: cryptographically signed C2PA manifests.
- What Is The Claude Watermark?An invisible, mandatory provenance mark on Claude output. Live since 2 August 2026.
- Claude Watermark Zero Width Characters: The MythClaude does not use zero-width characters. Scrubbing them leaves the real mark untouched.
What is covered
- Which Claude Models Have Watermarks?Launched on or after 2 August 2026: marked. Earlier models: being brought into scope.
- Does Claude Code Watermark My Code?In scope, yes. But low entropy plus formatters means little signal survives normal work.
- Does The Claude API Watermark Output?Yes. API output is marked, and it flows through into whatever you build on top.
Policy and consequences
- Can You Opt Out Of The Claude Watermark?No opt-out exists at any tier. The obligation sits on the provider, not on your preference.
- Claude Watermark And EU AI Act Article 50A provider-side transparency obligation, implemented globally, with deployer duties left to you.
- Claude Watermark False Positives When ProofreadingMarking tracks processing, not authorship. Proofread your own writing and it comes back marked.
Comparisons
Check what your text and files actually carry
Runs in your browser. No signup, no upload, nothing stored.