llms.txt
An llms.txt file sits at the root of a domain and describes the site in plain language for AI crawlers: what the company does, which pages matter, and how the content is organised. Unlike a sitemap, which lists URLs and nothing else, it is written to be read.
The proposal is genuinely new and adoption across major platforms is still uneven — it is worth being honest that publishing one is not a guaranteed win today. What it costs is an hour; what it offers is the chance to state your own positioning in your own words, rather than leaving a model to infer it from fragments of page copy.
That framing is the real argument for it. Everything else in AEO is indirect: you publish content and hope a system draws the right conclusion. This is the one file whose entire purpose is telling that system directly who you are.
AI crawlers have become an audience distinct from human visitors, with different needs and no ability to ask follow-up questions. llms.txt is currently the only mechanism for addressing them directly. Adoption is still early, but the cost of publishing one is an hour and the downside is nothing.
A company publishes an llms.txt summarising its product in three sentences, listing its glossary, pricing and docs, and defining the five terms central to its category. A crawler reading it gets an accurate orientation immediately, instead of assembling one from a homepage written for humans in a hurry.
What is llms.txt and do I actually need one?
llms.txt is a plain text file at your site's root that tells AI crawlers which pages matter and how to read them. Adoption is early, but it costs an hour.
