March 1, 2026
The robots.txt file tells search engines what to crawl and what not to. The llms.txt file does something similar, but for AI. It tells language models who you are, what you do, and where to find key information about your business.
In an era where an increasing share of web traffic is generated by AI crawlers and chatbots, it's no longer enough to have a website optimized only for Google. You need language models to understand your site too, and that's exactly what llms.txt is for.
AI models like ChatGPT, Claude, or Gemini can't interpret your website like a human. They don't see animations, don't understand context menus, and struggle to navigate complex HTML. Modern websites are full of JavaScript, dynamically loaded content, and visual elements that are invisible to AI.
Imagine your website has a beautiful animated landing page with minimal text. A human immediately understands what it's about. But an AI model only sees empty HTML with a few lines of JavaScript. It doesn't know if you sell shoes, build websites, or repair cars.
The llms.txt file is a simple text file in the root directory of your website that provides AI with a structured overview. It's like a business card for your website: clear, concise, and well-organized.
It's important to note that llms.txt is not a replacement for your website content. It's a supplement that helps AI understand the overall context and quickly orient itself in who you are and what you offer.
The format is simple: plain text with Markdown formatting. It starts with a heading with the name, followed by a description and sections with links. The specification was proposed by llmstxt.org and is already used by hundreds of websites including major tech companies.
The file should be written in the language of your target market. If you target both domestic and international markets, create two versions, for example, llms.txt in your primary language and en/llms.txt in English.
Here's a universal template you can use as a starting point for your own file:
# Company Name
> Short description of the company or project in one sentence.
More detailed description: what you do, for whom,
where you operate, and how you differ from competitors.
Can be 2–3 paragraphs.
## Services
- [Service Name 1](https://example.com/service-1): Short description of the service
- [Service Name 2](https://example.com/service-2): Short description of the service
## Portfolio / References
- [Project 1](https://example.com/portfolio/project-1): Project description
- [Project 2](https://example.com/portfolio/project-2): Project description
## Blog / Knowledge
- [Article 1](https://example.com/blog/article-1): What the article is about
- [Article 2](https://example.com/blog/article-2): What the article is about
## Contact
- Web: https://example.com
- Email: info@example.com
- Phone: +1 234 567 890
- Location: New York, USAEach section starts with a level-two heading (##). Links are written in the format [text](URL): description. The blockquote (>) at the beginning serves as a summary for quick AI orientation.
The llmstxt.org specification distinguishes two files:
Why have both versions? Some AI models have limited context windows and need a quick overview. Others process large volumes of text and appreciate the more detailed version. By offering both, you cover both scenarios.
In practice, I recommend starting with llms.txt and adding llms-full.txt later, once you have enough quality content on your site. There's no point in creating a detailed version if your website only has basic information.
The llms.txt file is still a relatively new format, but adoption is growing rapidly. These AI systems and their crawlers actively look for and process llms.txt:
Importantly, even crawlers without explicit llms.txt support can benefit from it. Since it's plain text in the root directory, any crawler can easily find and process it.
To let AI crawlers know where to find your llms.txt, you can reference it directly from robots.txt. It works similarly to a sitemap reference:
User-agent: *
Allow: /
Sitemap: https://example.com/sitemap.xml
LLMs-Txt: https://example.com/llms.txtThe LLMs-Txt directive isn't officially standardized like Sitemap yet, but several crawlers already recognize it. You can also link to llms.txt from your sitemap.xml, which increases the probability that crawlers will find it.
Another option is to add a link in your HTML <head>:
<link rel="llms-txt" href="/llms.txt">This meta tag helps crawlers that parse HTML headers. The more paths to llms.txt you provide, the better.
After creating your llms.txt, it's important to verify that it's properly formatted and accessible. Here are several methods:
https://yourdomain.com/llms.txt in your browser. It should display clean text without errors./llms.txt from AI crawlers (GPTBot, ClaudeBot, PerplexityBot).I recommend performing these checks at least once a month, especially after major website changes. Your llms.txt should always reflect the current state of your services and offerings.
The llms.txt format is rapidly gaining popularity. Among the early adopters are primarily technology companies and SaaS platforms:
But it's not just big companies. More and more small business owners, freelancers, and local businesses are adding llms.txt to their websites. The advantage for smaller companies is enormous: AI models can correctly categorize them and recommend them in answers to relevant queries.
Create a llms.txt file in the root directory of your website (next to robots.txt and sitemap.xml). Include basic information about your company, links to key pages, and a short description of each.
Follow these steps:
/llms.txt.LLMs-Txt directive to your robots.txt.The entire process takes 15–30 minutes. If you have a more complex website with dozens of pages, expect about an hour of work.
As part of GEO optimization, I create llms.txt for all websites I build. From my experience, results become visible within a few weeks of deployment.
For one client, a photographer based in Brno, we added llms.txt describing their services, specialization (weddings, portraits, corporate photography) and locations where they operate. Within a month, they started being recommended in ChatGPT answers to queries like "wedding photographer Brno." Previously, AI didn't mention them at all.
Another example: an e-commerce store selling Czech food products targeting international customers. We created an English llms.txt describing the product range, delivery terms, and unique products. AI chatbots started recommending it to foreign customers looking for Czech specialties.
Key finding: llms.txt works best in combination with other GEO techniques, including structured data, quality content, and authority signals. The file itself is important, but it's not the only step toward visibility in AI search engines.
Check out my llms.txt as an example. And if you need help, get in touch. I'll be happy to help set it up for your website.
Related articles: What is GEO Optimization • Why a Fast Website Makes More Money