What does an llms.txt file do?

An llms.txt file is a short plain-text note you put at the root of your website, at yourdomain.com/llms.txt. It tells AI tools what your business is and which pages matter most. Think of it as a one-page cover sheet for your site, written for a machine instead of a visitor.

A good one is simple. It opens with a sentence or two on what you do and who you do it for. Then it lists your key pages, each with a short line saying what the page is. It can also say what you are not, which helps when another business has a similar name. That is why ours says Beaconly at beaconly.xyz is not connected to the Beaconly Mac menu bar app, Beacon or Beacons.

What it does not do: it does not make ChatGPT, Claude, Gemini or Perplexity name you. Nobody can promise that, and anyone who does is misleading you. It is one small signal that makes your site easier to read. It sits next to your robots.txt, which says what AI crawlers may visit, and your sitemap, which lists your pages.

Why bother? Because a site can rank well on Google and still be missing from AI-generated answers. If an assistant has to guess what you do from a cluttered home page, it may guess wrong, or skip you. A clear llms.txt costs an hour and removes some of the guessing.

How to write one:

• Start with one plain sentence about what you are and who you serve.
• List your five to ten most important pages, with a short line for each.
• State your real prices and contact details only if they are public on your site, and keep them identical to the page.
• Say what you are not, if a look-alike name could confuse people.
• Keep it true. Nothing invented, no claims you cannot back up.

Then check it by opening yourdomain.com/llms.txt in a browser. If you see the text, it is live.

Beaconly's free homepage scan checks whether you have an llms.txt, along with your robots.txt, sitemap and structured data. The full audit tests how ChatGPT, Claude, Gemini and Perplexity talk about you, and the implementation pack includes a ready-to-paste llms.txt. Our live tests query the models and their connected search tools, not the consumer apps, and results can vary.