llms.txt
What is llms.txt?
llms.txt is a text file placed in the root directory of a website, analogous to robots.txt, but designed for language models (LLMs). It contains structured information about the site — who you are, what you offer, what content you have — in a format that is easy for AI to parse.
The llms.txt standard was proposed as a way to help AI models understand a site's context without having to scan hundreds of subpages.
Why does it matter?
- Faster understanding — AI gets a "cheat sheet" about your site instead of scanning hundreds of URLs
- Narrative control — you decide what information AI encounters first
- GEO — the llms.txt file is one of the foundations of AI optimization
- Growing adoption — an increasing number of AI tools check llms.txt before crawling a site
- Easy implementation — it is a plain .txt file that requires no special technology
How to create llms.txt?
File structure
# Company Name
> Brief description of the business (1-2 sentences)
## Company Data
- Name, address, tax ID
- Email, phone
- Website URL
## Services
### Service Name
URL: https://...
Service description
## Knowledge Base
### Article Title
URL: https://...
Description
## Contact
- Email, phone
Variants
- llms.txt — abbreviated version (key information)
- llms-full.txt — full version with a complete content inventory
Best practices
- Place it in the root directory —
/llms.txtand/llms-full.txt - Add to
<head>—<link rel="alternate" type="text/plain" href="/llms.txt"> - Update automatically — generate from your CMS/build process, not manually
- Markdown structure — headings, lists, URLs — a format readable by AI
- Include key facts — company data, services, metrics, unique value propositions
- Do not hide weaknesses — AI can distinguish marketing from facts
Details in the article llms.txt — what it is and how to create it.
Related Terms
- GEO — optimization for AI engines
- Robots.txt — access control file for crawlers
- Citability — content citability
- RAG — AI retrieval and generation mechanism
- AI Overviews — AI-generated answers in Google