LLMs.txt: What It Is, How to Create It, and Whether It Helps AI Visibility

LLMs.txt is a plain-text file you put at the root of your site to give AI models a clean, curated map of your most important content. Claude and Perplexity read it today. Google says it does not. So it is worth adding, as long as you know what it can and cannot do.

The short version: llms.txt is low-cost hygiene that helps on some AI engines and is ignored by others. It is not a ranking factor, and it does not replace the work that moves citations. This guide covers what it is, how to create one in a few minutes, and the honest evidence on whether it moves the needle in 2026.

Key Takeaways

• LLMs.txt is a markdown file at yourdomain.com/llms.txt that points AI models to your key pages, proposed by Answer.AI in 2024.
• It is confirmed read by Anthropic's Claude and by Perplexity, observable in ChatGPT's retrieval, and explicitly not used by Google.
• It is different from robots.txt (which controls crawler access) and your sitemap (which lists every URL for search crawlers).
• Adoption is still low, roughly one in ten sites, so publishing one is a cheap way to be ahead of most competitors on the engines that do read it.
• Treat it as hygiene, not a growth lever. Structured content, schema, and third-party authority still do the heavy lifting for AI citations.

What Is LLMs.txt?

LLMs.txt is a markdown file placed at your site's root that gives large language models a curated guide to your most important content. You publish it at yourdomain.com/llms.txt, and AI systems that support it can read it to understand what your site is about and which pages matter.

The format was proposed by Answer.AI in 2024 and has become the de facto standard for this. The idea is simple. Search crawlers are good at working through a full site. AI models work with limited context and benefit from a short, human-curated shortlist instead of the whole thing.

A basic llms.txt is readable markdown: an H1 with your brand name, a one-line summary in a blockquote, and a few sections of links to your key pages, each with a short description. An optional companion file, llms-full.txt, includes the full text of those pages inline for models that want the content directly.

Think of it as a hand-drawn map to your best rooms, not a floor plan of the entire building.

How LLMs.txt Differs From robots.txt and Your Sitemap

LLMs.txt does a different job from the two files it gets confused with. All three sit at your root, but they answer different questions.

• robots.txt controls access. It tells crawlers, including AI crawlers, which paths they may and may not fetch. It is permission, not curation.
• sitemap.xml is a complete inventory. It lists every URL you want search engines to know about, for machines, in XML.
• llms.txt is a curated shortlist. It is human-readable markdown that points AI models to the handful of pages you most want them to understand and cite.

Note: llms.txt does not grant or block access. If you need to keep AI crawlers out of a section, that is a robots.txt and server-level job, not something llms.txt can do.

How to Create an LLMs.txt File

You can write a working llms.txt in a few minutes. Here is the structure that matches the standard.

1. Start with an H1 and a summary. Put your brand or site name as the H1, then a single blockquote line describing what you do.
2. Group your key pages into sections. Use H2 headings like ## Docs, ## Guides, or ## Product, and under each, a markdown list of links.
3. Describe each link. After each link, add a short phrase on what the page covers. The description is what helps a model decide the page is relevant.
4. Keep it curated. List your best 10 to 30 pages, not everything. The whole point is to prioritize, so a bloated file defeats it.
5. Publish it at the root. Upload it so it resolves at yourdomain.com/llms.txt, then confirm it loads in a browser.

If you would rather not hand-write it, our free LLMs.txt generator builds a valid file from your site in about a minute, and you can edit the shortlist before you publish.

Does LLMs.txt Help AI Visibility?

Here is the honest answer: it helps on some AI engines and does nothing on others, and the effect is modest. The value depends entirely on which platforms your buyers use.

Who reads it today:

• Claude (Anthropic) and Perplexity publicly confirm they retrieve and use llms.txt, and both use it to help decide which of your pages to pull into an answer, per current adoption research.
• ChatGPT is not officially confirmed, but its retrieval patterns show correlated changes when a clean llms.txt is present.
• Google has said on the record that it does not use llms.txt, and Gemini's behavior does not change when you add one. Google's stack stays robots.txt, sitemaps, and structured data.

So llms.txt is genuinely useful if your category shows up on Claude and Perplexity, and irrelevant to your Google AI Overviews presence.

What the evidence shows on impact: Brands that publish a well-curated llms.txt see a modest but measurable uplift in citation rates, strongest on Anthropic and Perplexity, and strongest for sites with sprawling navigation that benefit from explicit curation. The gain is real but small.

Why it is still worth doing: Adoption remains low, roughly one in ten sites, so a clean file is a cheap way to be ahead of most competitors on the engines that read it. It costs you 30 minutes and carries no downside.

Here is how to think about it. If your buyers live in Claude and ChatGPT rather than Google, a tight llms.txt pointing to your docs and comparison pages helps those engines pick your best pages when they answer questions in your category. You will not see a traffic spike. What you can reasonably expect is your brand turning up a little more often in the exact place your buyers already look, for about 30 minutes of work. If your buyers live in Google, skip it and spend the time elsewhere.

Where it does not help: llms.txt will not fix thin content, missing schema, blocked AI crawlers, or a weak entity footprint. If AI is not citing you, the file is rarely the reason, and adding one will not be the fix. Think of it as the last 5% you add once the fundamentals are already in place.

FAQ

Is llms.txt a ranking factor? No. No AI platform treats it as a ranking or scoring signal. At most it helps a model that already supports it decide which of your pages to surface. That shapes which page a model picks. It does not change how much the model trusts you.

Do ChatGPT and Perplexity read it? Perplexity and Claude confirm they do. ChatGPT's use is observable but unconfirmed. Google says it does not use it at all. So the honest answer is "some do, some do not."

Where do I put the llms.txt file? At your site's root, so it resolves at yourdomain.com/llms.txt. An optional llms-full.txt at the same level holds the full page content inline for models that want it.

Does llms.txt replace my sitemap or schema markup? No. Your sitemap still serves search crawlers and your schema still helps every engine, including Google, parse your content. llms.txt sits alongside both.

Is it worth adding if adoption is only around 10%? Yes, for most brands. It takes 30 minutes, carries no risk, and the low adoption is the point: you get ahead of competitors on Claude and Perplexity while it is still uncommon.

Your Next Step

Generate a clean llms.txt for your site, publish it at the root, and keep the shortlist to your genuinely best pages. It is a 30-minute job that puts you ahead of 90% of sites on the engines that read it.

Then spend your real effort where citations are won. Start with a free AI SEO audit to see which of your pages are cite-worthy today, or read the complete AI SEO guide for the work that moves the needle across every engine, not just the two that read a text file.