Every few years a file appears that you “have to have”: once robots.txt, then sitemap.xml, today llms.txt. The question is whether it genuinely helps AI visibility or is just a fashion. Short answer: it is a cheap, harmless and sensible piece of the foundation — provided you do not treat it as a magic switch. The longer answer is below, together with our own file as an example.
What llms.txt is
llms.txt is a proposed simple standard: a Markdown text file placed in the root directory of a website (e.g. yourcompany.com/llms.txt) that describes in a dozen or so lines what the site is and points to the most important pages with a short explanation of each. The idea is analogous to robots.txt: instead of forcing a language model to wade through navigation, scripts and decoration, you give it a concise, ordered description — “what is here and where to find it”.
The structure is deliberately minimal:
- a heading with the company or site name,
- a one- or two-sentence summary in a blockquote,
- a short paragraph of context (what you do, for whom, key facts),
- sections with link lists: services, important pages, resources — each link with a one-sentence description.
Who reads it — honestly
Caution is needed here. llms.txt is a proposal, not an official standard, and none of the major assistant providers has publicly declared that it fetches the file systematically for every answer. So there is no proof that the file alone will “boost your visibility in ChatGPT”. What is certain: the file is publicly available, easy to fetch by any AI tool that browses your site, and harms neither SEO nor users. In our practice we treat it like structured data — a foundation element that lowers the cost of understanding your site for a machine.
What our file looks like
We do not ask anyone to do something we have not done ourselves. Our file lives at invisio.digital/llms.txt. It starts with the name and a one-sentence summary, then a paragraph with the facts we care about — since when we have operated, where we are from, what our starting prices are — followed by a list of services and important pages, each with a description. The file is generated automatically from the same data that builds the site navigation, so it never drifts away from the site’s content. That matters more than it seems: an llms.txt that says something different from the website is worse than none at all.
How to prepare your own — in five steps
- Collect the facts you care about. Name, since when you have operated, where, for whom, starting prices (if you publish them), contact. These are the sentences you want the model to “remember”.
- Pick a dozen or so key pages. Services, pricing, “about”, contact, your best articles. Not all of them — the file is a map, not a copy of the sitemap.
- Write one-sentence link descriptions. Specific ones: “WooCommerce stores from PLN 7,900 net with integrations”, not “Learn more”.
- Save as Markdown and upload to the root directory. Check that it opens at
/llms.txtas plain text. - Wire it into updates. Ideally the file should be generated from the same data as the navigation. If that is impossible — add “update llms.txt” to your website change checklist.
What llms.txt will not do
- It will not replace facts on the website. If a service page gives neither price nor scope, the model will still cite the competitor who does — the file cannot make up for it.
- It will not fix mismatches. Different years in business in an industry directory, a different rating in a widget than in Google Business Profile — the assistant will still see the contradiction.
- It does not work as advertising. Writing “the best agency in Poland” into the file achieves nothing; descriptions must be factual, because only those will the model treat as credible.
- It does not exempt you from indexing. Google AI Overviews and ChatGPT’s search rely on search-engine indexes, so classic technical SEO still matters.
Verdict
Worth it — because the cost is an hour of work, the risk is zero and the benefit is potential. But in the order of actions llms.txt is point four, not point one. First clear answers to customer questions on the website, then consistent facts everywhere the company appears, then Google Business Profile and reviews — and only then llms.txt and structured data to close the foundation. If you want to check what the picture of your company looks like in AI answers today, start with the free AI visibility audit — and read what the whole job involves in AI search optimization — what it is and how it differs from SEO.