What is llms.txt
llms.txt is a text file in the site root that tells large language models which information from the site can be used in answers. The idea is similar to robots.txt: robots.txt tells search crawlers what to index, while llms.txt tells AI crawlers which materials are most important.
The file was proposed as an open standard in 2024. Not every model reads it yet, but the trend is developing — Google, Anthropic and OpenAI are experimenting with the format.
Why llms.txt matters for business
Three practical tasks it solves:
- Directs AI crawlers to important content instead of hunting for relevant pages across the entire site.
- Provides a short brand description in machine-readable form — reduces the risk of entity confusion.
- Lets you highlight key pages — improves the chance of being cited.
How the file is structured
llms.txt is simple markdown. Minimal structure:
- H1 with the site or brand name.
- Short project description as blockquote.
- H2 sections with key links — services, cases, blog, contact.
- Each link as markdown: name, URL, short note.
What to write in each block
Brand description
One or two lines: who you are, what you do, which markets you serve. No marketing fluff.
Services
A list of key services with links. Service name — the real one.
Key pages
About, cases, blog, contact. What helps the LLM understand your site structure.
Languages and markets
For multilingual sites — list all versions and their geo.
Where to place the file
At the domain root at site.com/llms.txt. Served as plain text without auth.
How to verify the file works
Ask the model: 'What does site.com/llms.txt say about company X.' If it cites your description, the file is indexed.
Should you do it now
Yes. Takes an hour to write. Even if a specific model doesn't read it yet, you have a ready artifact that will start working as the format gains support.
