Guides
Understand AI watermarks, compare the evidence, and keep a clear record of your writing.
Provider watermark guides
Claude AI watermarks: what they show
Does Claude watermark text? Explore supported models, how the signal works, detector access, and what a saved draft can show readers.
OpenAI watermarks: ChatGPT, Codex and the API
Understand OpenAI’s textGrain watermark, the EU rollout, optional API marking, and the limits of text verification.
xAI and Grok watermarks: the announced API rollout
Explore xAI’s announced Grok text watermark rollout and opt-out controls, and why Overwride has opted out in favor of its own ledger.
Compare with Overwride
Overwride vs Claude watermarking
Compare Claude’s model watermark with Overwride’s draft records: reader access, writing history, authorship limits, and useful workflows.
Overwride vs OpenAI watermarking
Compare OpenAI’s textGrain signal with a shareable Overwride record. Understand verification, human contribution, and draft visibility.
Overwride vs xAI watermarking
Overwride has opted out of Grok text watermarking and uses its own ledger. Compare provider signals with shareable drafts, rewrites, and disclosure.
Build with the API
Writing and disclosure
How to disclose AI use in a post
Plain ways to say you used AI to write a post, with example lines for LinkedIn, newsletters and blogs, and a way to let readers check.
The EU AI Act and AI-assisted text
Since 2 August 2026 the EU AI Act has transparency rules for AI-generated text. What they cover, the editorial review exception, and how to keep a record.
Why AI detectors are a poor test for writers
AI detectors guess, and they guess wrong in ways that hurt real writers. What the evidence shows, and what to do instead if you use AI to write.