Does your site have a working llms.txt?

A markdown file at the root of your domain (/llms.txt) that tells AI assistants what your site is and where the useful pages are. It was proposed by Jeremy Howard in September 2024 and is being adopted as a de-facto convention. Think of it as a curated map for models, the way robots.txt is a set of rules for crawlers and a sitemap is an index for search engines.

Per the spec at llmstxt.org, exactly one thing is required: an H1 with the name of the site. Everything else is optional but useful - a blockquote summary right after it, then H2 sections each holding a markdown list of links in the form "- [Name](URL): note". A section named "Optional" marks links a model can skip when context is tight.

No, and they do not replace each other. robots.txt says which crawlers may fetch what - it is permission. llms.txt says what your site is and which pages matter - it is orientation. A site can allow every AI crawler and still be badly understood, which is exactly the gap this file closes.

Adoption is uneven and no major assistant has committed to it publicly, so treat it as cheap insurance rather than a guarantee. It costs one small file, it cannot hurt your search rankings, and if a model does look for it you are the site that has one. We would rather tell you that plainly than oversell it.

Because a map that points at dead pages is worse than no map. We sample a few of the links in your file and check they resolve. We only read public pages and never submit forms.

Read it first. Every link comes from your own sitemap and the name and summary from your own page metadata, so nothing in it is invented. Where we could not find a value we left a marked gap in <<< >>> for you to fill rather than guess. Once it reads right, save it as llms.txt at the root of your domain.