Definition
What is llms.txt?
llms.txt is a proposed standard: a plain-text, markdown-formatted file placed at the root of a domain (at /llms.txt) that gives large language models a curated, human-written map of the site's most important content. It is a hand-picked index with short descriptions and links, designed to be read directly by a model rather than crawled. It is emphatically not a robots.txt for AI — it grants nothing and blocks nothing.
Built by our community of 2,000 marketers
Free skills and prompts for paid ads and SEO
Templates for Claude, ChatGPT and Perplexity.
Clients we work withClients we
work with
llms.txt
llms.txt is a proposed standard: a plain-text, markdown-formatted file placed at the root of a domain (at /llms.txt) that gives large language models a curated, human-written map of the site's most important content. It is a hand-picked index with short descriptions and links, designed to be read directly by a model rather than crawled. It is emphatically not a robots.txt for AI — it grants nothing and blocks nothing.
Also written as: llms.txt file, LLM txt.
What the file contains
The format is deliberately minimal. An H1 with the site or organisation name, an optional blockquote summarising what the site is, then H2 sections grouping links, each link followed by a one-line description of what the reader will find there. That is essentially the whole specification.
The point of that austerity is that the file is written for consumption, not crawling. A model reading a sitemap gets several thousand URLs with no signal about which matter. A model reading a good llms.txt gets twenty URLs with a sentence explaining each — enough to decide what to fetch without burning context on the rest.
Some sites also publish per-page markdown versions of key documents, and an /llms-full.txt containing the full text of the core content rather than only links. That variant is more common in software documentation than in ecommerce, where the full catalogue would be far too large to inline usefully.
What llms.txt is not
Three misconceptions are worth clearing up, because all three circulate widely and each leads to wasted effort.
It is not an access-control file. robots.txt tells crawlers what they may fetch; llms.txt does not gate anything. Publishing one does not grant AI systems permission they lacked, and omitting a page from it does not hide that page.
It is not an official standard adopted by the major AI companies. It is a community proposal with meaningful voluntary adoption, particularly among developer-tool and documentation sites, and increasingly shipped by Shopify SEO apps. Whether any given assistant fetches it at request time is not publicly guaranteed, and you should treat vendor claims about it with appropriate scepticism.
It is not a ranking factor and it will not, on its own, get you cited. A file pointing at thin content simply helps a model find thin content faster.
Should an ecommerce store publish one?
The honest answer is: probably yes, because the cost is close to zero, but expect nothing dramatic. It takes an afternoon, it cannot hurt, and if adoption grows you are already positioned. Treat it as cheap insurance rather than as a growth lever.
If you do publish one, curate it rather than dumping your sitemap into it. The value is entirely in the selection. For a store, that usually means your buying guides, sizing and compatibility pages, comparison pages, shipping and returns policy, and your main collection pages — not every SKU. Write the one-line descriptions as if explaining the page to a colleague, because that is closer to how a model will use them.
Then keep it current. A stale llms.txt full of 404s is worse than no file, because it teaches whatever fetches it that your site is unreliable. If you are auditing your AI-search setup more broadly, the file is one small part of the structured data and AI visibility picture rather than the centre of it.
Frequently asked questions
Is llms.txt an official standard?
No. It is a community proposal with growing voluntary adoption, particularly among documentation and developer-tool sites, and it is increasingly shipped as a feature by Shopify SEO apps. It has not been formally adopted by the major AI companies, and no assistant publicly guarantees that it fetches the file at request time. Publish it as cheap insurance, not as a dependency.
Is llms.txt the same as robots.txt for AI?
No, and this is the most common misunderstanding. robots.txt is an access-control file that tells crawlers what they may and may not fetch. llms.txt grants nothing and blocks nothing — it is a curated index that helps a model find your most useful content quickly. If you want to control AI crawler access, that is still a robots.txt job.
Will adding llms.txt improve my rankings?
It is not a ranking factor and there is no evidence it moves classic search positions. Its plausible benefit is narrower: making it easier for a model to locate and correctly summarise your most important pages. A file that points at thin content just helps a model find thin content faster, so the underlying pages still have to be worth citing.
