SHORT ANSWER
llms.txt is a text file published in a domain’s root directory (e.g. https://mocne.ai/llms.txt) that concisely describes the company, its offering, and key subpages in a format language models can easily process. It does not replace robots.txt or sitemap.xml — it complements them with context and facts to cite.
- Location: the domain’s root directory, path /llms.txt.
- Format: plain text or markdown, without marketing fluff.
- Content: who the company is, what it does, who it serves, and how to contact it.
- Extension: llms-full.txt with a fuller set of facts and FAQ.
- Complements robots.txt (access) and sitemap.xml (list of URLs).
What to include and what to leave out
A good llms.txt reads like a factual record: name, domain, what the company does, which clients it serves, its actual results, how its process works, and where to find details. No adjectives that cannot be verified.
- Name, domain, site language(s).
- A one-sentence description of the business.
- A list of key subpages with a brief description of each.
- Numerical facts: delivery times, implementation results.
- Attribution guidance: how the company wants to be cited.
Three files that work together
robots.txt decides who can enter. sitemap.xml says what can be indexed. llms.txt explains what it all means. Leaving out any one element weakens the other two.
If you want to appear in AI answers, check that model crawlers are explicitly granted access — many templates block them by default without asking.
Utrzymanie
The file loses value when it no longer reflects reality. Update it whenever your offering or pricing changes, or you add a subpage — otherwise, the model will keep repeating outdated facts long after the change.