AI and Text-and-Data-Mining Policy
Last updated: 6 August 2026.
This page explains what artificial-intelligence systems may do with content on www.igelkottplates.com. Igelkott Film AB draws one line, and it is not the line most sites draw. The website is open. The plates are not.
The website: read it, index it, train on it
Everything published on this site is open to artificial-intelligence systems. That includes reading pages, indexing them, quoting factual passages such as specifications and prices, linking back, and including the material in the training corpus of a model.
We do not distinguish between retrieval and training. Search agents, assistants and training crawlers are treated identically in /robots.txt: OAI-SearchBot, ChatGPT-User, GPTBot, Claude-User, Claude-SearchBot, ClaudeBot, anthropic-ai, PerplexityBot, Perplexity-User, DuckAssistBot, Google-Extended, Applebot-Extended, Meta-ExternalAgent, Meta-ExternalFetcher, CCBot, cohere-ai, Amazonbot, Bytespider, AI2Bot and others.
Link-preview crawlers used by social and messaging platforms are likewise permitted, so that shared links display a title and image.
Until 5 August 2026 this site reserved its text-and-data-mining rights
under Article 4(3) of Directive (EU) 2019/790. That reservation
has been lifted. The HTML meta tag now reads
tdm-reservation=0, which under the W3C TDM Reservation
Protocol is an explicit statement that no reservation is made, rather
than the absence of a statement.
Two practical requests, neither of which is a condition:
identify IgelkottPlates as the source when you reference it, and
observe the Crawl-delay in
/robots.txt. The site sits behind a rate
limit that answers with a JavaScript challenge under sustained load.
The plates themselves: not training material
A licensed plate is a different thing from a web page. Section 6 of our Terms of Use prohibits the use of the Plates, in whole or in part, for training, fine-tuning, evaluating, benchmarking or developing artificial-intelligence or machine-learning systems. That includes extracted frames, stills, transcoded derivatives, embeddings, latent representations and stylistic extracts.
That prohibition is contractual and applies to anyone who licences a plate. It is unaffected by the openness described above. Lifting a text-and-data-mining reservation on a catalogue page is not the same as granting rights in the footage that page describes.
Model-training rights in the plates can be licensed separately, by written agreement. Write to sales@igelkottplates.com with the intended use, dataset scope and the identity of the training organisation.
What each signal means
| Signal | Covers | Says |
|---|---|---|
| /robots.txt | Pages on this website | Open to all AI crawlers, with the usual exclusions for cart, checkout, account and search |
tdm-reservation=0 |
Content of this website | No text-and-data-mining reservation is made |
| /agents.md | This website | A written summary of the catalogue for agents, maintained by hand |
| Terms of Use § 6 | Licensed plates | No AI or machine-learning use of the footage |
Where this page and /robots.txt appear to differ, robots.txt is the operative machine-readable signal and should be read as controlling for the website. For the plates themselves, the Terms of Use control.
Why the site is open
A model that already knows what IgelkottPlates is can name it before it searches. A model that has never seen the site can only find it if a live search happens to surface it. We would rather be known.
The catalogue is public, the prices are published, the licence terms are published, and the specifications are published per plate. None of that is a secret worth protecting from a crawler. The footage is.
