SEO has always been a game of optimization. We optimized titles, we optimized links, we optimized speed. Now, we must optimize rights.

Text and Data Mining (TDM) rights are the new battleground. As Large Language Models (LLMs) hunger for training data, they must navigate a minefield of copyright law. The EU’s DSM Directive explicitly allows TDM exceptions unless rights are “expressly reserved” by the rights holder in a machine-readable format.

That format is TDMREP.

The Hybrid Strategy: Public Face, Private Brain

Many SEOs make the mistake of seeing TDMREP as an all-or-nothing switch. “Block AI!” or “Allow AI!” This binary thinking is obsolete. The optimal strategy is a hybrid model.

1. The Public Face (Allow Training):

   • Content: “About Us”, Product Documentation, Press Releases, Basic definitions.
   • Strategy: Allow full TDM rights. Why? You want the AI models to “know” who you are. You want your brand to be part of their parametric memory. If you block training on your “About Us” page, you risk the model hallucinating your history.

2. The Private Brain (Reserve Rights):

   • Content: Original research, deep analysis case studies, proprietary datasets, premium news.
   • Strategy: Use TDMREP to signal “Reservation.” This tells the crawler: “You can index this for search (blue links), but you cannot ingest this for training.”

Monetizing the Authorization Layer

This reservation is not just defensive; it is the first step in monetization. By asserting your rights via TDMREP, you create a “Licensing API” for your content. When an enterprise crawler (like OpenAI’s or Anthropic’s) hits a reserved page, it can be programmed to check for a licensing endpoint.

If you have a partnership deal in place, the crawler authenticates, bypasses the reservation, and ingests the premium content. This turns TDMREP into a dynamic paywall for AI.

Implementation Details

The technical implementation is straightforward but critical. You must serve the reservation in a way that is machine-readable and legally binding.

The most robust method is via the HTTP Header:

X-TDM-REP: /tdmrep.json

This points to a JSON file at the root of your domain (e.g., https://mcp-seo.com/tdmrep.json). Inside this JSON, you define your policies using the W3C vocabulary.

For example:

{
  "policy": "https://mcp-seo.com/licenses/tdm-reservation",
  "technique": "http://www.w3.org/ns/tdmrep#text-and-data-mining",
  "target": [
    {
      "path": "/premium-research/*",
      "action": "reserve"
    },
    {
      "path": "/public-docs/*",
      "action": "allow"
    }
  ]
}

Conclusion

By implementing a granular TDMREP strategy, you move from being a passive source of training data to an active participant in the data economy. You protect your assets while ensuring your brand remains visible in the AI’s world model. This is the definition of Agentic SEO: controlling the narrative not just in the SERPs, but in the neurons of the models themselves.