AI Search Readiness guide · Chapter 5
llms.txt: what it is and what it is not
A markdown map of your site for agents and documentation tools. Useful, cheap, and not a ranking signal. Here is how to write one honestly.
What the file is
llms.txt is a proposal for a markdown file at the site root: a heading with the site name, a one-line summary, and lists of links to the pages that explain the site, each with a short description. Coding agents and documentation tools read it; it is not confirmed as an input to ChatGPT search, Google AI features or Perplexity ranking, and the vendors have not said it is.
The audit treats it as one low-severity check inside the Meaning layer. A missing file costs a little; a present file with a heading, a summary and at least three live same-origin links earns the points back. It can never turn a blocked Access or Receipt layer into a ready one.
Writing it
Start with "# Site name", then "> one sentence on what the site is and who it is for", then a "## Pages" list linking pricing, about, docs, features, FAQ and contact with a few words each. Keep every link live and on your own domain; agents follow them. Update it when pages move.
The questions layer of the audit drafts one from the pages that exist, and the repair writes it only when the file is missing, linking to paths it has verified are live.
Serving it
Some hosts return the HTML shell for unknown paths with a 200 status, which makes a missing llms.txt look present. The audit fetches the file and checks the body is text, not HTML. Serve it as text/plain or text/markdown from the public folder.