llms.txt: How to Help ChatGPT, Claude and Perplexity Cite Your Site Correctly
In 2026, AI crawlers look for llms.txt, a file in the site root that explains to models what your site is about and which sources are canonical. ChatGPT, Claude and Perplexity already read it. Most Russian-language sites don't have one, and models cite them poorly or not at all. The file takes about 30 minutes to write, and the effect on citation rates shows up within 1–4 weeks.
AI-processed from Habr AI; edited by Hamidun News
In 2026, visibility in the answers of ChatGPT, Perplexity and Claude is no longer a privilege of large publishers. It is a necessity for every site that wants to stay relevant. The problem is that AI crawlers often pick up incomplete or distorted information about you. They rely on general knowledge from the training set rather than pulling data straight from your site. llms.txt addresses exactly this: it is a simple text file in the site root that tells language models who you are, what you do and how to cite you correctly.
How llms.txt works
llms.txt resembles robots.txt but works in the opposite direction. robots.txt controls how ordinary crawlers (Googlebot, Yandexbot) scan a site, telling them which pages to crawl and which to ignore. llms.txt is an instruction for the language models themselves: when they generate an answer to a user's query, they check whether the site has this file, and if it does, they follow your guidance on citation and sources.
When a user asks ChatGPT or Claude about your company, the model can look into llms.txt and get current information about who you are, what you do and how to cite you. This matters especially because the training data of large models is updated rarely (often once a year or less), so the information goes stale while you pivot, change your services or launch a new product.
Most Russian sites don't create this file. As a result, models generate answers from generalized knowledge in the training set, often misrepresenting your position, mixing you up with competitors or not mentioning you at all. With llms.txt you take explicit control over how you are presented in AI answers.
In 2026, more than 60% of online information searches already start with ChatGPT or Perplexity rather than Google. If a user asks a model about your line of business and no llms.txt exists, they will get either outdated information or a blend of data about your competitors. That is a direct risk of losing customers and of your market position being misunderstood.
The 5 blocks it should contain
A minimal llms.txt should contain five sections:
- Description — a one-line description of your project (who you are, who you work for, what you write about)
- Full description — an extended explanation of your mission, target audience and examples of work (3–5 paragraphs)
- URL mapping — a list of the site's key sections with a short explanation (what is in the blog, contacts, offers)
- Requirements — exactly how models should cite you (whether a link is required, attribution format, citation style)
- CDN URLs — if your media files live on separate domains (images.example.ru, video.example.ru), list them here
This is the minimum. Later you can add file versioning, content license information, recommendations on update frequency or a list of main authors.
Example for a production site
Here is what it looks like in practice:
Description: Hamidun.ru — a blog about AI for engineers and startup founders
Full description: We look at how modern language models work, how to use them in production and how to build AI into a company's business processes. The target audience is developers, technical leads and founders who want to understand the current state of AI and find practical uses for it in their projects.
URL mapping: /blog/news — fresh news and announcements from the AI world /blog/tools — reviews and comparisons of AI tools /blog/deep-dives — detailed breakdowns of model architecture and real-world cases /contacts — feedback form
Requirements: Cite hamidun.ru as the primary source, attach a hyperlink to the specific article, and name the author where one is given
CDN URLs: images.hamidun.ru, media.hamidun.ru
Last updated: 2026-05-21
Upload the file to the root of your domain (next to robots.txt and sitemap.xml). Models usually find and start using updates within 1–4 weeks. The first effect on citation in AI answers appears after roughly a week and stabilizes by the fourth week.
What this means
llms.txt evens the odds between large information resources and small projects. Before, a small site simply dissolved in the training context of large models. Now you can state explicitly: "Here is my content, cite exactly this, and here is how to do it correctly." It is slower than organic traffic from Google, but the result is more stable: models follow your instructions rather than generating approximations from random information.
The main thing is not to put it off. llms.txt takes 30 minutes to write, and the result works for years. Every day without the file is a missed chance to be cited correctly in AI answers, a channel that is growing faster than Google.
- Nov 19, 2026Anthropic goes public and raises $81.3 billion, beating SpaceX's $75 billion record
- Nov 19, 2026Infostart connects Direct, Metrica and Wordstat to ChatGPT: 6 commands, shared access
- Nov 19, 2026Amazon and Google's Anthropic Stakes Valued at $308 Billion in IPO at $83 a Share
- Nov 19, 2026Google's Gemini 4 Neon runs twice as fast as Carbon and costs 17% less per task
Key Questions Answered
How llms.txt works?
llms.txt resembles robots.txt but works in the opposite direction. robots.txt controls how ordinary crawlers (Googlebot, Yandexbot) scan a site, telling them which pages to crawl and which to ignore. llms.txt is an instruction for the language models themselves: when they generate an answer to a user's query, they check whether the site has this file, and if it does, they follow your guidance on citation and sources.
What are the key details on The 5 blocks it should contain?
A minimal llms.txt should contain five sections: Description — a one-line description of your project (who you are, who you work for, what you write about) Full description — an extended explanation of your mission, target audience and examples of work (3–5 paragraphs) URL mapping — a list of the site's key sections with a short explanation (what is in the blog, contacts, offers) Requirements — exactly how models should cite you (whether a link is required, attribution format, citation style)…
What are the key details on Example for a production site?
Here is what it looks like in practice: Description: Hamidun.ru — a blog about AI for engineers and startup founders Full description: We look at how modern language models work, how to use them in production and how to build AI into a company's business processes. The target audience is developers, technical leads and founders who want to understand the current state of AI and find practical uses for it in their projects.
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.