What Is llms.txt and Why Your Company Needs It in 2026
llms.txt is a new standard for communicating with AI crawlers. We explain what it is, how to write it properly, and why companies that implement it first gain a lasting advantage.
Let's start with an analogy that makes this instantly clear.
You probably know robots.txt — the file that tells Google what it may and may not index. llms.txt is its equivalent for the AI age. But while robots.txt gives instructions, llms.txt gives context.
What is llms.txt
llms.txt is a text file placed in the root directory of a website (yourwebsite.com/llms.txt). It was proposed in 2024 by the community around Jeremy Howard as a standardized way for websites to communicate with large language models.
Unlike robots.txt, which controls access, llms.txt provides AI systems with direct, structured information about:
- What your company does
- Who is behind it
- What its key pages and content are
- How to correctly refer to the company and its products
It's like handing AI systems a company brochure in a format they can process efficiently.
Why it works
LLMs process the web continuously — indexing content and building an "understanding" of companies and concepts. But the web is full of unstructured data. The same information can be phrased in a thousand different ways across a thousand different pages.
llms.txt solves this problem simply: you give the AI exactly the information you want it to have. Properly written, in a clear structure, with the key facts in one place.
The result: AI systems have a more accurate, more consistent, and more complete understanding of your company — which translates into more accurate and more frequent citations.
How to write llms.txt properly
The format is intentionally simple — readable text with Markdown structure. The key sections:
Header with identification:
# Company Name - llms.txt
# One-sentence description
# Web: https://yourwebsite.com
# Contact: email@yourwebsite.com
About section — who you are, what you do, for whom, and what makes you different. Write it as if you were explaining your company to an AI assistant in a single paragraph.
Services or products section — a concrete offering with prices, timeframes, and deliverables. Structured data is easier for AI to process.
Key facts section — statistics, founder, year founded, area of operation, languages.
Site structure section — the main URLs with descriptions.
Contact — direct contact information.
The ideal length of an llms.txt is 500–2,000 words. Too short and it won't provide enough context. Too long and AI systems will skip over it.
llms.txt vs llms-full.txt
The standard defines two files:
llms.txt — a concise version for quick loading. AI systems use it for basic context and orientation.
llms-full.txt — the complete version with a detailed description of all services, methodology, FAQ, case studies, and any other content you want available for deeper AI analysis.
Both files should be accessible without authentication and served correctly with Content-Type: text/plain; charset=utf-8.
Who currently supports llms.txt
As of late 2025, llms.txt is actively used by:
- Perplexity AI — explicitly mentions it as an indexing signal
- Obsidian, FastAI, Anthropic — as early adopters within their ecosystems
- Hundreds of SaaS and tech companies globally
ChatGPT and Gemini don't have publicly confirmed direct support yet, but both systems index accessible text content — so a properly written llms.txt works as part of the site's overall AI context.
Where llms.txt stands now: the first-mover advantage
In the Czech Republic, an estimated less than 1% of websites have llms.txt implemented. That's an exceptional situation: a low barrier to entry, but a high potential payoff.
Companies that implement llms.txt today are in a better position for a simple reason: AI models are trained continuously on data from the web. Content present in the training data shapes how the model understands a company. The sooner structured information is available, the more solid the foundation for the AI's representation of your brand.
Practical implementation steps
- Create a text file
llms.txtin your website's root directory - Write structured information about your company (see above)
- Add an
llms-full.txtwith the complete version - Verify the files are available at
yourwebsite.com/llms.txt - Add explicit permissions for AI crawlers to your
robots.txt - Test accessibility via a browser or curl
For a typical website, the whole implementation takes no more than 2–4 hours. The effect on AI visibility shows up at the next indexing — typically within weeks.
An important caveat
llms.txt is a tool, not a miracle. The file alone, without quality content on the website, isn't enough. AI systems combine the information in llms.txt with what they find on your other pages. If the information doesn't match, the model will notice.
llms.txt works best as part of an overall GEO strategy — alongside citable content, proper schema markup, and a correctly configured robots.txt.
Frequently asked questions
What is the llms.txt file? A Markdown-formatted text file placed at a website's root (domain.com/llms.txt) that provides AI language models with a structured summary of the company: what it does, its services, prices, and key pages.
Is llms.txt an official standard? Not yet — it's an emerging convention that AI systems are gradually adopting. Implementation takes minimal time, carries no risk, and gives you a first-mover advantage while your competitors don't have it.
What is the difference between llms.txt and llms-full.txt? llms.txt is a concise index with links and a summary; llms-full.txt contains the website's complete content in a single file for models that want to load everything at once.
Is llms.txt enough to be visible in AI? No — it's a complement, not a miracle. It has to be consistent with your website's content and work alongside citable content, schema markup, and a properly configured robots.txt.
Need help implementing llms.txt as part of a comprehensive GEO strategy? Get in touch — as part of our GEO Audit + Strategy, we implement and configure it for you.
Founder of the first GEO studio in the Czech Republic. He helps companies become an authority for ChatGPT, Gemini and Perplexity.
Related articles
Schema markup for GEO: Which types matter most and how to implement them
A guide to schema markup from a GEO optimization perspective. FAQPage, Organization, Article, HowTo - what to implement first and why schema markup directly affects AI citability.
robots.txt for AI Crawlers: The Complete Guide to GPTBot, ClaudeBot, and PerplexityBot
How to properly configure robots.txt for AI crawlers. Who GPTBot, ClaudeBot, PerplexityBot, and Google-Extended are, and how to make sure they can access your content.