What is llms.txt and what should it contain?
We request /llms.txt at the root and check that it returns 200 with content an agent can read.
What it is
llms.txt is a single Markdown file served at the root of your domain, at /llms.txt. It tells a language model or an agent, in plain prose, what your organisation does and which pages matter.
It is not robots.txt and it is not a sitemap. robots.txt grants or refuses access, a sitemap lists every URL without saying which are important, and llms.txt is the short human-written summary that gives an agent orientation before it starts fetching.
The convention is a first-level heading with your name, a blockquote summarising what you sell, then second-level headings grouping linked pages with a one-line description each.
Why an agent needs it
An agent arriving cold has to infer your business from whatever HTML it happens to fetch first, which on most marketing sites is a hero image and a navigation menu.
llms.txt replaces that inference with a statement. It is the cheapest file on this list to publish and the one that most changes what an agent understands about you.
The Fix
Publish a Markdown file at /llms.txt served as text/plain or text/markdown. Open with `# Your company name`, then a blockquote of one or two sentences saying what you sell and to whom.
Group the pages that matter under second-level headings — Services, Pricing, Contact — as Markdown links with a short description after each. List the pages a buyer would need, not every page you have.
Keep it under a few kilobytes and update it when your offer changes.
A Deep Scan writes this one for you. It arrives in the Fix Pack, generated from the pages we crawled on your site.