agenticweb.wiki

TDMRep (TDM Reservation Protocol)

Last checked · 25 primary sources

This page is about the TDM Reservation Protocol (TDMRep) of the W3C Text and Data Mining Reservation Protocol Community Group: a way for rightsholders to say that their text-and-data-mining rights are reserved, as an opt-out under Article 4 of the EU DSM Directive, and to point to a licensing policy [1][10]. It is not the IETF AIPREF work (AIPREF vocabulary, AIPREF Content-Usage attachment), Cloudflare's Content Signals Policy, Really Simple Licensing (RSL), Spawning's ai.txt, or AI crawler user-agent tokens in robots.txt, which have their own pages.

Overview

TDMRep lets a rightsholder state in machine-readable form that text-and-data-mining rights over web content are reserved, which is how the spec expresses the opt-out from the text-and-data-mining exception in the EU copyright directive [1][10]. The same two properties can sit in a JSON file under /.well-known, in HTTP response headers, in HTML meta elements, or inside EPUB and PDF metadata, and the spec sets rules for which one a mining agent should check first [1]. An optional policy link, written as an ODRL profile, tells miners how to contact the rightsholder or get a licence [1]. The group describes the flag as a reservation of rights rather than a ban on mining, and nothing in the protocol itself forces an agent to comply [1][4].

How it works

Who uses it

Each site below was fetched without following redirects on 2026-10-29 (sweep run 02:00–02:01 EET, re-run in the self-check). Method: 118 sites (49 named in the Community Group's adopters list [8] and 69 other publishers and news sites) were checked for /.well-known/tdmrep.json and for TDMRep headers and meta tags on the home page. A file was counted when it parsed as a JSON array of rules with location and tdm-reservation, after at most one redirect to another tdmrep.json, which the spec allows [1]; headers and meta tags were counted only on home pages that returned 200. Across two runs (02:00 and 02:20 EET), 51–52 of 118 sites showed at least one signal: 37 valid tdmrep.json files, 5 sites with TDMRep headers and 14–15 with TDMRep meta tags (some sites use more than one); the difference is Éditions du Seuil, whose home page was blocked in one run. 33 of the 49 listed adopters and 18–19 of the 69 other sites showed a signal. Many sites block automated requests, so these counts are a floor. Other counts: EDRLab says "Of the top 250 French websites, 143 have already implemented the TDMRep specification." (undated) [18], and a count posted to the group's GitHub on 2026-09-17 is titled "Deployment data: TDMRep declarations on 28 of 7,282 observed hosts (frozen 10,667-host sample, Sep 2026)" [21].

Who reads it

No statement from a major AI model developer or crawler operator that it reads TDMRep was found. Lantad reported: "Read on 4 August 2026, the crawler documentation published by OpenAI, Anthropic and Google names robots.txt in every case and mentions none of the four." [19] EasyxLab archived crawler documentation on 2026-10-03: "We could read it for 8 of 14 providers: OpenAI, Anthropic, Google, Mistral AI, Apple, Amazon, Perplexity and Common Crawl." "None of the 8 names TDMRep, Content-Signal, the IETF Content-Usage or noai." [20] The only reader found is Spawning's datadiligence filtering library and API.

Recent coverage

Background: the Commission says of the GPAI Code of Practice: "The code was published on July 10, 2025." [13] Its Measure 1.3(1)(b) asks signatories "to identify and comply with other appropriate machine-readable protocols to express rights reservations pursuant to Article 4(3) of Directive (EU) 2019/790, for example through asset-based or location-based metadata" without naming a protocol [12]. The Commission's consultation FAQ lists TDMRep first: "In addition to robots.txt, the study identifies seven other technical solutions (TDM reservation protocol (TDMRep); C2PA TDM Assertions; AI.txt; …" [11].

See also

References

  1. W3C Text and Data Mining Reservation Protocol Community Group, TDM Reservation Protocol (TDMRep), Final Community Group Report, version 3, 2024-05-10: "Final Community Group Report 10 May 2024" "Editor: Laurent Le Meur (EDRLab)" "Previous version TDMRep, Second Version" "published by the Text and Data Mining Reservation Protocol Community Group under the W3C Community Final Specification Agreement (FSA)." "It is not a W3C Standard nor is it on the W3C Standards Track." "This initiative is a technical answer to the constraints set by the Article 4 of the new European Directive on copyright and related rights in the Digital Single Market." "This specification defines a JSON file named tdmrep.json, which MUST be hosted in the /.well-known repository of a Web server." "In each rule, location and tdm-reservation are mandatory, tdm-policy is optional." "tdm-reservation and the optional tdm-policy are two properties added to the HTTP header of the response to a GET or HEAD request." "This is currently the preferred technique for implementing the protocol, as it is simple and already integrated in the swpawning.ai API." "tdm-reservation is expressed as value of the name attribute of a meta element and tdm-policy is expressed as value of the name attribute of a second meta element." "They both use the tdm prefix, which is a shorthand mapping for the http://www.w3.org/ns/tdmrep# URL." "tdm:reservation and the optional tdm:policy are two elements added to the XMP metadata section of the PDF." "A PDF/A validator will raise errors if TDMRep properties are present in the XMP metadata section of a PDF document." "A TDM Agent MUST check the presence of a TDM file on the origin server before it starts scraping the content of the Web server." "The format of policies defined in this specification is a profile of the Open Digital Rights Language 2.2 [ODRL]." "The "opt-out" option specified by the Article 4 of the CDSM Directive is expressed by the use of tdm-reservation with value equal 1." — https://www.w3.org/community/reports/tdmrep/CG-FINAL-tdmrep-20240510/
  2. W3C TDMRep Community Group, TDM Reservation Protocol, Final Community Group Report, version 2, 2024-02-02: "Final Community Group Report 02 February 2024" — https://www.w3.org/community/reports/tdmrep/CG-FINAL-tdmrep-20240202/
  3. W3C TDMRep Community Group, TDM Reservation Protocol, Final Community Group Report, version 1, 2022-02-16: "Final Community Group Report 16 February 2022" — https://www.w3.org/community/reports/tdmrep/CG-FINAL-tdmrep-20220216/
  4. W3C, Text and Data Mining Reservation Protocol Community Group home page (final reports register and posts), checked 2026-10-29: "2024-08-09 TDM Reservation Protocol, version 3" "2024-04-16 TDM Reservation Protocol, version 2" "2022-02-16 TDM Reservation Protocol, version 1" "Notes, September 30th, 2025 Laurent Le Meur | Posted on: October 1, 2025" "It is not a NO-TDM signal, but a RIGHTS-RESERVED signal." — https://www.w3.org/community/tdmrep/
  5. W3C Team, Call for Participation in TDM Reservation Protocol Community Group, 2021-01-20: "Posted on: January 20, 2021" "This group was originally proposed on 2021-01-18 by Laurent Le Meur." — https://www.w3.org/community/tdmrep/2021/01/20/call-for-participation-in-tdm-reservation-protocol-community-group/
  6. W3C, Participants in the Text and Data Mining Reservation Protocol Community Group, checked 2026-10-29: "Chair: Laurent Le Meur EDRLab" "Chair: Giulia Marangoni AIE - Italian Publishers Association" — https://www.w3.org/community/tdmrep/participants
  7. GitHub API, repository metadata for w3c-cg/tdm-reservation-protocol, checked 2026-10-29: ""full_name": "w3c-cg/tdm-reservation-protocol"" ""pushed_at": "2026-09-03T14:00:40Z"" — https://api.github.com/repositories/331841798
  8. W3C TDMRep Community Group, docs/adopters.md (GitHub, main branch), checked 2026-10-29: "This page contains a list of adopters of the TDM Reservation Protocol." "Elsevier (https://www.elsevier.com/): as tdmrep.json, http headers + html metadata, with a policy." — https://raw.githubusercontent.com/w3c-cg/tdm-reservation-protocol/main/docs/adopters.md
  9. IANA, Well-Known URIs registry (CSV), checked 2026-10-29: "tdmrep.json,[https://www.w3.org/2022/tdmrep/],provisional,W3C,,2024-02-15," — https://www.iana.org/assignments/well-known-uris/well-known-uris-1.csv
  10. Directive (EU) 2019/790 on copyright and related rights in the Digital Single Market, Article 4(3), Official Journal L 130, 17.5.2019, Publications Office (Cellar) copy of the OJ text: "The exception or limitation provided for in paragraph 1 shall apply on condition that the use of works and other subject matter referred to in that paragraph has not been expressly reserved by their rightholders in an appropriate manner, such as machine-readable means in the case of content made publicly available online." — https://publications.europa.eu/resource/cellar/214471fe-786e-11e9-9f05-01aa75ed71a1.0006.03/DOC_1
  11. European Commission, Stakeholder consultation on AI and copyright compliance (FAQ), last updated 2025-12-01: "In addition to robots.txt, the study identifies seven other technical solutions (TDM reservation protocol (TDMRep); C2PA TDM Assertions; AI.txt;" "The Copyright Chapter of the GPAI Code includes Measure 1.3 (Identify and comply with rights reservations when crawling the World Wide Web)." — https://digital-strategy.ec.europa.eu/en/faqs/stakeholder-consultation-ai-and-copyright-compliance
  12. European Commission, General-Purpose AI Code of Practice, Copyright chapter (PDF), Measure 1.3(1)(b): "to identify and comply with other appropriate machine-readable protocols to express rights reservations pursuant to Article 4(3) of Directive (EU) 2019/790, for example through asset-based or location-based metadata," — https://ec.europa.eu/newsroom/dae/redirection/document/118115
  13. European Commission, The General-Purpose AI Code of Practice, checked 2026-10-29: "The code was published on July 10, 2025." — https://digital-strategy.ec.europa.eu/en/policies/contents-code-gpai
  14. W3C, TDMRep Community Group meeting at TPAC 2025 (event page), meeting on 2025-11-14: "We'll then discuss a possible evolution of TDMRep, using the Vocabulary defined by AI Pref." — https://www.w3.org/events/meetings/4d16ca76-dc36-4bf2-914e-13c2ca89ad2d/
  15. GitHub API, comments on rom1504/img2dataset issue #308 (comment by Padge91 of Spawning, 2023-07-11 UTC): "`datadiligence` currently supports the existing noai HTTP headers, the proposed TDMRep HTTP headers, and the Spawning API, with additional methods planned." — https://api.github.com/repos/rom1504/img2dataset/issues/308/comments
  16. Spawning, datadiligence documentation, datadiligence.rules module, checked 2026-10-29: "This class wraps logic to evaluate the TDM Reservation Protocol headers:" — https://datadiligence.readthedocs.io/en/latest/api/datadiligence.rules.html
  17. Spawning, datadiligence documentation, datadiligence.evaluators module, checked 2026-10-29: "HttpEvaluator (user_agent = None, respect_robots = True, respect_tdmrep = True)" — https://datadiligence.readthedocs.io/en/latest/api/datadiligence.evaluators.html
  18. EDRLab, TDM Reservation Protocol (EDRLab open standards page, undated), checked 2026-10-29: "Of the top 250 French websites, 143 have already implemented the TDMRep specification." "The adoption of TDMRep by TDM/AI actors is currently indirect: Spawning AI has implemented the protocol in an API used by Stability.ai and others." — https://www.edrlab.org/open-standards/tdmrep/
  19. Lantad, Three crawler documentation pages name robots.txt, and none names TDMRep, 2026-08-04: "Read on 4 August 2026, the crawler documentation published by OpenAI, Anthropic and Google names robots.txt in every case and mentions none of the four." — https://lantad.co/blog/tdmrep-not-named-in-crawler-documentation
  20. EasyxLab, Text-and-data-mining reservations in the EU-27 press, channel by channel (Study S16, working draft), published 2026-10-03: "We could read it for 8 of 14 providers: OpenAI, Anthropic, Google, Mistral AI, Apple, Amazon, Perplexity and Common Crawl." "None of the 8 names TDMRep, Content-Signal, the IETF Content-Usage or noai." — https://easybyte.es/lab/studies/s16/paper/
  21. GitHub API, w3c-cg/tdm-reservation-protocol issue #62 (opened by flober81, 2026-09-17 UTC): ""title": "Deployment data: TDMRep declarations on 28 of 7,282 observed hosts (frozen 10,667-host sample, Sep 2026)"" — https://api.github.com/repos/w3c-cg/tdm-reservation-protocol/issues/62
  22. European Commission, AI Office, Summary report of the stakeholder consultation supporting Article 53(1)(c) AI Act and Measure 1.3 of the GPAI Code of Practice (PDF, undated; copy linked from CEPIC's post [23]; no Commission-hosted copy was found): "For rightsholders, TDMRep is the most widely deployed rights reservation protocol from those surveyed, although experience and evaluation of the tool depend on the sector, with the protocol being mostly deployed within the books/press publishing sector." "Responses from GPAI model providers reveal predominantly critical or cautious assessments of the TDMRep protocol." "A number of providers state that the protocol has limited real-world adoption and uneven uptake across sectors." — https://drive.usercontent.google.com/download?id=1I9VKaxjZniu9iT-2Xzmw-mQcbzftkzOb&export=download
  23. CEPIC, European Commission Publishes Summary Report Following TDM Opt-Out Protocol Workshop, 2026-06-10: "The European Commission has published the Summary Report following the stakeholder workshop on Text and Data Mining (TDM) opt-out protocols held on 2 June 2026" — https://www.cepic.org/post/european-commission-publishes-summary-report-following-tdm-opt-out-protocol-workshop
  24. STM Association, EU Commission's consultation on TDM rights reservation protocols, 2026-06-15: "A first workshop took place on 2 June, with TDMRep emerging as the protocol with the most support among rightsholders." "While AI companies have pushed back, the possibility of it being made mandatory increases its relevance." — https://stm-assoc.org/eu-commissions-consultation-on-tdm-rights-reservation-protocols/
  25. Boniface de Champris, Copyright & AI: Why the European Commission Should Avoid Flawed Opt-Outs and Support International Standardisation (Disruptive Competition Project), 2026-09-02: "Developed by French publishers via a W3C community group, rather than a formal standards body, TDMRep is not a W3C standard and lacks balanced multistakeholder governance." "In practice, TDMRep's location-based controls offer no additional functionality over Robots.txt." — https://project-disco.org/european-union/copyright-ai-why-the-european-commission-should-avoid-flawed-opt-outs-and-support-international-standardisation/

JSON · Markdown · OKF v0.2