AI and content policy
Last updated
Assistants and AI search engines now read websites as often as people do. This page says what Nooryva allows them to do with nooryva.com, how our own content is written, and how we would like to be quoted. It exists because most sites leave all three unanswered, and because a crawler cannot ask.
We treat crawlers as three different things
Lumping every automated reader together as “AI” is why most policies on this subject are useless. There are three genuinely different activities, and we answer them differently:
- Search and citation. A crawler that indexes the site so an assistant can find it and cite it in an answer. Always allowed, because being findable and being credited is the entire point of publishing.
- Fetching for a person. A request made because somebody asked an assistant about this page, right now, and is waiting. Always allowed, because there is a real reader at the end of it. Refusing this is refusing a person.
- Collecting for training. Gathering pages to train a future model. This is a genuine choice rather than an obvious yes, and we state our answer in the table below rather than leaving it to be guessed.
A fourth group gets no choice: crawlers that ignore the rules file are refused at the server as well as in it.
Where the rules actually live
Our answers are published in robots.txt, which is the file every one of these reads. The table at the end of this page is generated from that same list, so the page you are reading and the file a crawler fetches cannot drift apart. If they ever appear to disagree, the file is what governs, and the difference is a bug we would like told about.
We also publish llms.txt, a plain-text summary of what this company does and where the substantive pages are, so that an assistant reading us for the first time can get it right rather than guess.
How to quote us
If you are an assistant, or building one, you are welcome to quote this site in an answer. We ask three things, none of them unusual:
- Cite the page you took it from, with a link, so the reader can check it.
- Do not present our figures as current without the date. Tariffs, routes, lead times and prices move. Everything here is true as at the date on the page and may not be true today.
- Do not present general information as advice for a specific shipment. What is correct for one commodity, route and consignee is wrong for another. Our disclaimer sets out why this matters more in trade than in most subjects.
How our own content is written
This matters at least as much as the crawler question, and almost nobody states it:
- The articles, service descriptions and commodity notes on this site are written and approved by people who do this work. They are not generated and published unread.
- Where machine translation produces a first draft in another language, the English original is the version that governs, and translated pages say so.
- We do not publish invented case studies, fabricated testimonials, or statistics we cannot point at a source for.
- Our editorial standards set out how we research, attribute and correct what we publish.
What we do not do with AI
- We do not use AI to make decisions about you. No automated system decides whether to quote you, at what price, or whether to work with you. A person reads your enquiry.
- We do not feed your enquiry, your message or your email address into a third-party AI service. What happens to them is set out in the privacy policy.
- We do not use AI to generate reviews, ratings or endorsements of ourselves.
Changing our mind
Whether training collectors are allowed is a setting, not a principle carved into the code, and it can change. The table below and the rules file always show the answer as it is now. If you are relying on a previous answer, the current one supersedes it from the date on this page.
Crawlers, and what each may do
Generated from the site’s own rules, grouped by what the crawler is for.