Free tool: runs in your browser
AI text watermark writing record checklistwhen you use an AI tool to draft, translate or proofread text for a client, court, publisher or school
TimeBinder adds a layer of protection against AI and human fraud and disputes. Check which tools touched which passages, what the receiver’s disclosure rules say, and whether you keep dated drafts.
Nothing you tick is sent anywhere. This page is not legal advice and cannot detect a mark in a text. It does not help you remove one. Each provider’s own pages are the authority on how its marking works.
Check how AI touched the text, the rules that apply and the drafts you keep
Tick what is true about the text. You get a shareable result and clear next steps. Nothing you tick is sent to TimeBinder.
Your choices stay on this page. Nothing you tick is sent anywhere or tied to your name.
Tick everything that is true about the text you are about to send
This looks at whether you have written down which tools touched which passages, whether you have read the disclosure rules of whoever receives the text, whether you are treating a mark or a detector result as proof, and whether you keep dated drafts. It is not legal advice and cannot detect a mark.
What to do now
AI tools now mark their text. What should I keep if I use one to write? The short answer
Write down which tools you used, for what, and which passages they touched, and keep the prompts, drafts and edit history with dates. Anthropic’s help page says supported Claude models weave an imperceptible watermark into generated text, that it may persist through some editing, and that detection is in private preview for eligible organisations rather than for the public. It also says a detected mark is a signal that content may have been processed by Claude, not full proof of provenance, because people use Claude to proofread, translate or summarise their own writing, and that a missing mark does not mean AI was not used.
This checklist is not legal advice. A TimeBind of a file you already hold, such as a draft or an edit history export, can show that exact file existed no later than a recorded time and still matches. It does not show who wrote a text or whether it carries a mark.
Why this came up in November 2026
Anthropic says on its help page on how Claude marks AI-generated content that it has signed the EU AI Act’s Article 50(2) Code of Practice on Transparency of AI-Generated Content, that Claude models launched in the EU on or after 2 August 2026 support machine-readable marking at launch, and that it is adding text watermarks to earlier models. The page lists limits: marks may not survive heavy editing, paraphrasing or translation, very short passages leave too little text for a reliable signal, and Sonnet 4.6 may not carry the mark with non-default sampling settings. Posts on X also report that other providers are adding marks. Check each provider’s own documentation, because this page does not verify them.
Measured attention is visible. Checked on 13 November 2026, about 7:10 AM Australia/Brisbane. One public post on X about Claude’s text watermarks, from 2 November, had 98,373 impressions and 628 likes, and another from 31 October had 4,943 impressions and 72 likes. Google autocomplete for “claude watermark” returned “claude watermark detector”, “claude watermark checker” and “claude watermark explained”. Autocomplete shows what people type. It is not search volume.
Many of those searches are about removing a mark. This checklist covers the other side: knowing what your tools did, following the disclosure rules of whoever receives the text, and keeping dated drafts. For files, see the Content Credentials checker.
Source: Claude Help Center, How Claude marks AI-generated content.
What this checklist does and does not do
The honest job of this page is to put the checks in a sensible order, not to decide who wrote a text. See also the AI citation filing checklist and the glossary entries for existed no later than and fingerprint.
| Part | What it does | What it does not do |
|---|---|---|
| How AI touched the text | Reminds you to write down the tools, what they did and which passages they touched. | Detect a mark in a text. |
| The rules | Reminds you to check the disclosure rules of the court, publisher, school, employer or client. | Tell you what a rule requires. |
| Your drafts | Reminds you to keep prompts, drafts and edit history with dates. | Prove who wrote a text. |
| A TimeBind of a file you already hold | Records that an exact file existed no later than a recorded time and still matches, using a SHA-256 fingerprint made in your browser. The file stays on your device. | Show who wrote a text, or whether a text carries an AI mark. |
If your drafts and prompts are something you may need to show later
Before you send the text:
- Keep dated copies of drafts, prompts and edit history for each text an AI tool helped with.
- Make a TimeBind of each exact file you already hold. The fingerprint is made in your browser and the file stays on your device. You can later download a Proof of Time Certificate.
- Store the files, your notes on the tools used and the certificate together, with the date of each export.
A TimeBind shows the exact file existed no later than the recorded time and still matches. It cannot show who wrote a text.
First Pooled TimeBind free, no card, then from US$5. An ASAP TimeBind is US$15.
What a TimeBind shows, and what it does not. It shows that an exact file existed no later than a recorded time and that a copy shown later still matches its fingerprint. It does not show who made the file, who owns it, that its contents are true, or whether AI was used to make or edit it. This is general information, not legal advice.
To record, stage by stage, what you wrote or chose yourself and what an AI tool produced, see also the free AI-assisted work human contribution log. It does not decide whether the work is protected by copyright.
AI text watermark writing record checklist: FAQ
Does Claude watermark its text?
Anthropic’s help page says supported Claude models embed an imperceptible watermark in generated text. It lists which models are supported and says detection is in private preview for eligible organisations.
Does a detected mark prove that AI wrote a text?
Not according to Anthropic’s page. It says a detected mark is a signal that content may have been processed by Claude, and that Claude may not be the original author, for example when it proofread or translated your own writing.
Does no mark mean no AI was used?
No. The same page says a missing mark does not mean AI was not used, for example when the text was heavily edited, paraphrased, translated or too short.
Should I remove a mark before I send a text?
Do not scrub or paraphrase text only to hide AI use where disclosure is required. Check the rules of whoever receives the text, and say what you used when they ask.
Does a dated record prove that I wrote a text?
No. A TimeBind of a file you already hold, such as a draft, can show that exact file existed no later than a recorded time and still matches. It does not show who wrote a text. This page is not legal advice.
Is anything I tick sent to TimeBinder?
No. The checklist runs in your browser. Your ticks stay on the page.
Related guides and tools
- AI citation filing checklistAI in court filings
- Assignment AI use checklistAI use in coursework
- Client file AI tool checklistClient files and AI tools
- Content Credentials checkerSee what a file’s credentials say
- GlossaryPlain definitions of fingerprint, SHA-256 and existed no later than
- All free toolsPrivate, in-browser tools
Date the exact files you already hold
Add a layer of protection against AI and human fraud and disputes: a dated, checkable record of your drafts, kept before you send the text.
- Open the app and sign in with email or Google.
- Add your files. They stay on your computer.
- Seal them and download your Proof of Time certificates.
First Pooled TimeBind free, no card, then from US$5. See pricing.