What is llms.txt?
llms.txt is a Markdown file published at the root of a site (/llms.txt) that gives language models a curated map of the content. It opens with a summary of what the site is, then lists the main pages with a one-line description each. Proposed in 2024 as a convention, it does for AI what the sitemap does for search engines, except it curates and adds context instead of dumping every URL.
What does an llms.txt look like?
Plain Markdown, written for a model to read:
# Site name > One-paragraph summary: what the site is, > who it serves and what it specializes in. ## Main pages - [Title](https://example.com/page): one-line description. - [Another page](https://example.com/other): one-line description.
This site's own file follows exactly that format: rebovisibility.com/llms.txt.
How does llms.txt compare to robots.txt and sitemap.xml?
| FILE | WHAT IT DOES | FORMAT |
|---|---|---|
| robots.txt | Says which bots may access what | User-agent / Allow directives |
| sitemap.xml | Lists every URL with its modification date | Exhaustive XML |
| llms.txt | Curates the main pages with context | Markdown that models can read |
Source: the llmstxt.org proposal; application: ReBo architecture, 2026.
Related entries
AI crawlers are what consume the file, and robots.txt has to let them through first. It ships as one of the deliverables in the publishing stage of the ReBo method and works alongside structured data in the technical foundation of AI Visibility.
Frequently asked questions
Does llms.txt replace sitemap.xml or robots.txt?
No, they are complementary. robots.txt says who may access the site, sitemap.xml lists every URL for indexing, and llms.txt curates the most important pages with natural-language context, in the format models read best (Markdown).
Do AIs officially read llms.txt yet?
Adoption is still a convention in the making, and not every provider has declared official support. It costs almost nothing to maintain and doubles as a canonical map of the site, which is why we recommend publishing it.
What goes into llms.txt?
An H1 title with the site name, a summary paragraph as a blockquote, and sections with lists of links to the main pages, each link carrying a one-line description. Only essential content, always up to date; it is curation, not a URL dump.