llms.txt is a Markdown file placed at the root of a domain that guides AI models in understanding the site. It describes who you are, what you do, and which content is worth reading, optimizing citability in generative systems like ChatGPT, Perplexity, and Claude.

What's the difference between llms.txt and robots.txt?

robots.txt tells crawlers what they can or can't read. llms.txt tells AI models what to read to understand the site. The former manages access, the latter manages the meaning and priority of content.

Does llms.txt improve visibility on ChatGPT and Perplexity?

Yes. A well-structured llms.txt file increases the likelihood that AI models correctly cite the site in their answers. It helps the model understand the domain's context and the author's authority.

How do you create an llms.txt file?

An llms.txt file is a Markdown document with sections for: who you are, what you offer, and your main content with links. It should be placed at the domain root (e.g., webeseo.it/llms.txt) and updated whenever the main content changes.

llms.txt — The New SEO for Artificial Intelligence

llms.txt: The File That Talks to AI

A Markdown file at the root of your domain. Simple. Structured. Built not for Google, but for ChatGPT, Gemini, Perplexity, Claude.

This is llms.txt.

The Difference from robots.txt

robots.txt tells crawlers what they can and can't crawl. It's an access filter.

llms.txt is different. It doesn't block. It guides. It tells language models which pages on the site are authoritative, where to find the main information, and what tone represents the brand.

An LLM crawling a site without llms.txt has to infer on its own what's important. It might draw on outdated pages, comments, secondary content. It might generate answers that are inaccurate or disconnected from the brand's current identity.

llms.txt solves this problem by giving the model an explicit starting point.

What It Looks Like

The format follows the Markdown specification. A short description of the site at the top. A list of links to the main pages. The most important guides. The about page. Technical documentation, if there is any.

The hierarchy is intentional: Markdown is LLMs' preferred language because it conveys structure without the noise of HTML. Every token counts.

There's also llms-full.txt, a version that aggregates the entire text content of the site into a single document. For models with large context windows, it's even more effective.

Why It Matters for Visibility

When a user asks a chatbot for information on a topic, the model builds its answer by drawing on the sources it recognizes as authoritative. Being that source is the goal of GEO.

llms.txt doesn't guarantee citation. But it increases the odds that the model correctly understands the brand's positioning, uses up-to-date information, and doesn't attribute incorrect information.

It's an E-E-A-T signal in a non-Google environment.

What to Do Now

Create the file at the root of the domain. Write a concise description of the site. Add links to the pillar pages, the main guides, the contact page.

If you already have an llms.txt, keep it updated whenever the site's structure changes significantly.

GEO isn't a discipline separate from SEO. It's a layer added on top of a solid technical foundation. llms.txt is the bridge between the two.