llms.txt
llms.txt is a proposed plain-text file, written in Markdown and served at a site's root, that gives language models a short description of the site and a curated list of links to its most useful pages. Jeremy Howard of Answer.AI proposed it at llmstxt.org in September 2024. It is a convention, not a standard that any major AI assistant has publicly committed to reading.
What the file contains
A title with the site name, a one-paragraph summary, then sections of links, each with a short note on what the page covers. The format is plain Markdown, so a person can read it as easily as a model. The proposal also suggests that sites offer a Markdown copy of each useful page at the same address with .md appended.
What it is not
It is not robots.txt. It grants or blocks no access, and a crawler that ignores it loses nothing. It is also not a ranking factor: nobody has shown that publishing one makes an assistant name or cite a site more often. Being cited still depends on pages that answer questions directly and on presence across the sources the assistants already trust.
Why publish one anyway
It is cheap, it cannot hurt, and it gives any tool or agent that does read it a clean map of the pages that matter. It also forces a useful decision about which pages best describe the product. CitePrism publishes its own at /llms.txt, generated from the same content as the site and the sitemap, so the index does not drift out of date.