TDM Reservation Protocol - the W3C signal that satisfies EU AI Act opt-out
Tests for the W3C TDMRep declaration that satisfies EU machine-readable opt-out requirements.
What this signal tests
We check whether your site implements the W3C Text and Data Mining Reservation Protocol (TDMRep). This can take three forms: a `tdm-reservation` HTTP response header, a `/.well-known/tdmrep.json` document, or a `<meta name="tdm-reservation">` tag. When `tdm-reservation=1` (reserved), the site must also point to a `tdm-policy` URL describing the terms under which mining is allowed.
Why it matters for your visibility in AI
TDMRep is the machine-readable opt-out mechanism the EU Digital Single Market Directive (Article 4) and the EU AI Act require. If you are a European publisher - or if you serve European users - TDMRep is the legally-relevant signal that says "we have opted out of text and data mining for AI training." It is the only mechanism a court can point to and say "the site declared its preference; the AI company was on notice." The practical consequence: if your business has any commercial interest in not having its content swept into model training, TDMRep is the single most defensible declaration to publish today. It is a W3C Community Group report (not a full Recommendation) but it has been adopted by major European publishers, citation networks, and copyright registries. The legal weight is real even if the technical adoption among crawlers is still building. For sites that want to allow training under licence, the tdm-policy URL gives you a clean way to advertise terms.
Pass criteria at a glance
| Criterion | Passes when |
|---|---|
| At least one channel declares tdm-reservation; if 1, tdm-policy URL provided. |
How we test it
We check three channels. First, GET /.well-known/tdmrep.json and validate against the TDMRep JSON schema (an array of objects with `location`, `tdm-reservation`, and optional `tdm-policy`). Second, HEAD a sample of your pages and look for `tdm-reservation` and `tdm-policy` response headers. Third, parse the HTML head of a sample page for `<meta name="tdm-reservation">`. We pass if at least one channel makes a valid declaration, and - when reservation is `1` - a tdm-policy URL is present.
Show technical detection method
(1) GET /.well-known/tdmrep.json validates against TDMRep schema. (2) HEAD pages for tdm-reservation/tdm-policy headers. (3) Parse <meta name=tdm-reservation>.
If your site fails: how to fix it
- Decide your stance: reservation=0 (mining permitted) or reservation=1 (mining requires a licence per your tdm-policy URL).
- If reservation=1, write a tdm-policy page on your site stating terms, pricing if any, and contact for licensing. This URL is what crawlers and courts will read.
- Publish /.well-known/tdmrep.json with content like `[{"location":"/","tdm-reservation":1,"tdm-policy":"https://yoursite.com/tdm-policy"}]`.
- Optionally mirror the same declaration in HTTP response headers (`tdm-reservation: 1`, `tdm-policy: https://...`) for crawlers that prefer per-resource signals.
- Read the W3C report at https://www.w3.org/community/reports/tdmrep/CG-FINAL-tdmrep-20240202/ before implementing - this is best handled by a developer who can map the spec to your stack.
Quick facts
| Maturity | EMERGING |
|---|---|
| Weight | medium |
| Category | Emerging Standards |
Primary sources
Related signals
Frequently asked questions
Is TDMRep a real standard?
It is a W3C Community Group final report, not a full W3C Recommendation. However, it is the most mature machine-readable opt-out mechanism aligned with EU copyright law, and it is cited explicitly in EU AI Act compliance guidance.
Which AI crawlers honour TDMRep today?
Adoption among crawlers is partial - some honour it, some do not. The bigger point is that TDMRep is the signal a European court will look for when assessing whether a site "reserved its rights" under Article 4 of the DSM Directive.
I'm not in Europe. Should I still implement this?
If you have any European users or revenue, yes. If you operate purely in jurisdictions without an equivalent regulation, the aipref draft (see the related Content-Usage signal) may give you broader coverage with less ceremony.
What's the difference between TDMRep and ai.txt or robots.txt?
TDMRep is legally anchored to EU copyright law; robots.txt is purely a voluntary crawler convention. ai.txt is a Spawning-led proposal with overlapping scope. TDMRep has the strongest legal standing of the three in Europe.
Run your own scan
Run a free scan and see how your site grades across all 155 AI-readiness signals.