llms.txt
llms.txt is a text file in a website’s root directory that gives large language models (LLMs) a compact, easy-to-read summary of the company and its most important pages.
The idea is similar to robots.txt: instead of rules for search engine crawlers, llms.txt contains a short description in Markdown — who the company is, what it offers and which pages contain the most important information.
It is an open proposal from 2024, not a binding standard; not every AI system reads the file. But it takes little effort to create and is a sensible complement to structured data and a company facts page.
Circle provides its own llms.txt — with company facts, services and links to the most important pages.
What to look out for
- Place the file as /llms.txt in the domain’s root directory.
- Start with a short description: who is the company, what does it offer, for whom?
- Link the most important pages — services, references, company facts, contact.
- Keep content concise and factual, without marketing speak.
- Update the file when services or company data change.
Common mistakes
- llms.txt is treated as a replacement for structured data or good content.
- The file contains details that differ from the legal notice and website.
- Hundreds of unweighted links instead of the truly important pages.
Frequently asked questions
No. robots.txt controls which areas crawlers may visit. llms.txt summarises content for language models. They serve different purposes.
Usually yes, because the effort is small. It complements structured data and a facts page — but the bigger effect comes from good, unambiguous content on the website itself.