What is llms.txt?

llms.txt is a plain-text (Markdown) file served at https://example.com/llms.txt. It gives language models and AI agents a short, authoritative description of a site: what the product or organisation is, who it is for, the main URLs and how to get in touch. It is the AI-agent equivalent of robots.txt plus an about page, written to be read in one pass without rendering HTML.

It is a convention rather than a standard, proposed in 2024 and adopted by documentation sites and SaaS products first. Answer engines and coding assistants read it when it exists, and it is the cheapest way to control the sentence an engine uses to describe you.

The format

One H1 with the name, a blockquote with a one-paragraph definition, then short sections with lists of links. Keep it under a few hundred lines; a separate llms-full.txt can carry long-form content.

Example record
Name
https://example.com/llms.txt
Type
text/plain
Value
# Example > Example is a scheduling tool for small clinics. It handles bookings, reminders and intake forms. ## Main pages - [Home](https://example.com/) - [Pricing](https://example.com/pricing) - [How reminders work](https://example.com/reminders) ## Contact - [email protected]

Writing the definition

The first sentence is the one that gets quoted. Use the definitional form: "<Brand> is <category> for <audience>. It does X, Y and Z." Name the category in words your customers use, not an invented one, and say only what the product does today. Put the same sentence in the first paragraph of your homepage and in the meta description so every extractor agrees.

Publishing it

  • Serve it at the root of the canonical host with HTTP 200 and Content-Type text/plain (text/markdown is also fine). A 200 that returns your HTML app shell is a soft 404 and counts as missing.
  • Static sites: drop the file in public/ or static/. Single-page apps: make sure the catch-all rewrite to index.html does not swallow it. WordPress: several SEO plugins now generate it; otherwise upload it to the web root.
  • Keep the links absolute and on the canonical host. Update it when pages move; a stale llms.txt is worse than none.
  • Do not put claims in it that are not on the site (customer counts, uptime numbers). Engines quote it verbatim.

Frequently asked questions

Do search engines use llms.txt?
Google and Bing have not announced ranking use. Its value is with answer engines, agents and assistants that read plain text, and as a single place where your own definition of the brand lives.
Is llms.txt the same as robots.txt?
No. robots.txt says which paths crawlers may fetch; llms.txt says what the site is and where the important pages are. They complement each other: robots.txt points crawlers at the sitemap, llms.txt gives models the summary.
How does AstraVerify check it?
It fetches /llms.txt on the canonical host and passes the check when it returns 200 with a text content type and at least a short body. Missing or HTML responses lose the 4 points in Answer-engine readiness and produce a drafted file built from your title, description and sitemap.
Do I need llms-full.txt too?
Only if you have long documentation worth exposing as one plain-text file. Most marketing sites need only llms.txt.

Check your own domain. The scan shows your live records, a score out of 100 and the exact record to publish for each fix.

Related guides

Canonical: https://astraverify.com/what-is-llms-txt