llms.txt
Also known as: llms-full.txt, /llms.txt
llms.txt is a Markdown file placed at a website's root that briefly tells large language models what the site is about and which pages matter most; it was proposed in 2024 and is not yet an official standard.
Jeremy Howard proposed llms.txt in September 2024, with the details published at llmstxt.org. The idea is that a language model shouldn’t have to dig through HTML full of menus, ads and scripts to understand a site; it can read a short, tidy summary instead.
What goes in the file
Under the proposal, the file lives at /llms.txt and is written in Markdown:
- The name of the site or project (an H1)
- A one- or two-sentence summary (a blockquote)
- Optional paragraphs with more detail
- Lists of links to key pages, grouped under H2 headings
Some sites also publish an llms-full.txt that puts all their content in a single file.
Does it actually work?
It’s worth being clear: llms.txt isn’t an official standard. None of the major AI companies has said clearly that its crawlers routinely read the file and use it in answers. Google representatives have said Google doesn’t use it for search. So far it’s mostly been adopted for software documentation, where it helps coding assistants get to grips with a library quickly.
It’s easy to create and does no harm, but don’t expect it to drive visibility on its own. What matters more is giving AI crawlers access through robots.txt and making sure the pages themselves are well written.
Example
A law firm’s llms.txt might contain the firm’s name, a two-sentence description, and links grouped under “Practice areas”, “Lawyers” and “Contact”. For more detail, see our guide to AI crawlers and llms.txt.