What is llms.txt? The complete guide for site owners

If you run a website, you have probably added a robots.txt file at some point, or at least know that one exists. It sits at the root of your domain and tells search engine crawlers where they can and cannot go. llms.txt is a newer idea built on the same instinct, aimed at a different audience: instead of guiding search crawlers, it hands AI tools a clean, readable summary of your site.

This guide explains what llms.txt actually is, what the file looks like, who reads it today, and whether it is worth publishing one in 2026. It is written to be genuinely complete, including the parts that are inconvenient for anyone selling a plugin. If you leave with a clear head about what this file does and does not do, it has done its job.

What is llms.txt? The one-sentence answer

llms.txt is a Markdown file at the root of your website (yourdomain.com/llms.txt) that gives large language models a curated, plain-text map of your most important content.

The analogy to robots.txt is useful but only goes so far. robots.txt is a set of access rules: it tells crawlers what they may fetch. llms.txt is the opposite posture. It is not about restriction, it is about presentation. It says, in effect, “here is my site, described cleanly, so an AI tool does not have to wade through navigation menus, cookie banners, and layout markup to understand what I offer.”

The standard was proposed by Jeremy Howard of Answer.AI in 2024. It is published and documented at llmstxt.org. Importantly, it is a community convention, not an official standard from the W3C, the IETF, or any formal standards body. That matters for setting expectations, and we will come back to it.

What does an llms.txt file look like?

Part of the appeal is how simple the format is. It is just Markdown, following a light structure. A minimal llms.txt looks like this:

# Acme Coffee Roasters

> Small-batch coffee roaster in Portland. We sell single-origin beans,
> subscriptions, and brewing equipment, and publish brewing guides.

## Products

- [Single-origin beans](https://acme.example/beans): Rotating selection of
  single-origin coffees, roasted weekly.
- [Subscriptions](https://acme.example/subscribe): Monthly bean deliveries.

## Guides

- [Brewing guides](https://acme.example/guides): How to brew pour-over,
  French press, and espresso at home.

## About

- [Our story](https://acme.example/about): Who we are and how we roast.

The structure is deliberately readable by a human and a machine at the same time. An H1 with the site name, an optional blockquote summary, then H2 sections that group links, each link followed by a short description of what it points to. There is no schema to validate against and no required fields beyond the title. If you can write a Markdown list, you can write an llms.txt.

llms.txt vs llms-full.txt: what is the difference?

You will often see a second file mentioned alongside it: llms-full.txt.

The difference is scope. llms.txt is the index: titles, links, and one-line descriptions. It is a table of contents. llms-full.txt is the same structure with the actual content of each page inlined directly into the file. Rather than linking out to your brewing guide, llms-full.txt contains the full text of that guide, converted to clean Markdown, right there in the document.

Why have both? An AI tool that just needs to understand your site’s shape can read the compact index. A tool that wants the substance without making dozens of follow-up requests can read the full file in one fetch. The index is fast and cheap; the full file is complete. Publishing both covers both cases.

For a large site, llms-full.txt can get big, so it is common to include your most important pages there rather than every URL. The index can be broader.

Who reads llms.txt today?

This is the question that matters, and it is where a lot of marketing copy quietly overstates things. Here is the honest picture as of 2026.

AI tools that read a site before acting on it use it. This is the real, live use case, and it is more relevant to a normal store or blog than it first sounds. The clearest example for you is a shopping or research agent: an AI tool that browses a store to answer “which of these wallets is the lightest?” or pulls up an article to summarize it. Rather than crawl your cluttered HTML, it can fetch your llms.txt and get a clean version of your content. This is not hypothetical for stores. Shopify built its entire agent-shopping push around exactly this behavior, which is why Shopify stores now automatically generate an llms.txt file for every store.

Autonomous agents that act on someone’s behalf use it. As agent-style features spread across AI assistants, a clean map of your site’s structure saves an agent from parsing your rendered HTML before it books, buys, or answers. This surface gets more useful as those features spread, not less.

Some in-product AI assistants are reported to fetch it. There are credible reports that assistants like Perplexity and Claude reference llms.txt in their retrieval workflows. Treat these as encouraging signals rather than guarantees; the companies have not all made firm public commitments, and behavior can change.

Does llms.txt help SEO or Google rankings?

Google Search does not use it. This is not speculation. In June 2025, Google’s John Mueller said publicly that no AI system was using llms.txt at that time. Google then went further: in a June 2026 update to its AI-optimization guidance, it stated plainly that llms.txt has no effect, positive or negative, on Search rankings or on AI Overviews. If your goal is Google ranking, this file does nothing for you.

The big consumer assistants do not use it for ranking or citations. No major AI company (Google, OpenAI, Anthropic, Meta, Mistral) has publicly committed to consuming llms.txt in production as a ranking or citation signal. There is no measurable evidence that publishing one gets you cited more often by ChatGPT, Claude, Perplexity, Gemini, and other AI assistants in their consumer answers.

And adoption on the reading side is genuinely low. An Ahrefs study of roughly 137,000 sites in 2026 found that about 97% of llms.txt files never received a single AI request. Files are being published far faster than they are being read.

You will also see llms.txt marketed under labels like GEO (generative engine optimization) or AEO (answer engine optimization). The goal those terms describe is real: as more people start their research inside an AI assistant rather than a search box, being easy for those systems to read and quote genuinely matters. But llms.txt is not a proven lever for it, and no study to date has found a correlation between publishing one and being cited more often. Anyone selling the file as an AI search ranking tactic is ahead of the evidence.

None of this means the file is useless. It means you should publish one for the right reasons.

Which sites already publish an llms.txt?

The standard started in developer circles. Documentation-heavy, API-first companies like Anthropic, Stripe, and Cloudflare were early to publish an llms.txt, because their audience lives in AI coding tools that read these files. That is real, but if you run a store or a blog, it is also easy to shrug off as “not my world.”

So here is the part that is your world. It has spread to ecommerce, and it has spread fast.

As of May 2026, Shopify generates an llms.txt automatically for stores on its platform, as part of its push into agent-driven shopping. Wix does the same for its premium ecommerce plans. That means millions of ordinary online stores now ship one without the owner lifting a finger. These are not developer tools. They are the shops you actually buy from.

You can open real examples right now:

• Ridge (ridge.com/llms.txt), the minimalist wallet brand, publishes a clean, detailed index of its collections, pricing, materials, shipping, and return policy, exactly the kind of thing a shopper might ask an AI assistant about.
• Allbirds (allbirds.com/llms.txt), the shoe brand, publishes one aimed at shopping agents that browse and buy on a customer’s behalf.

Both of those are Shopify stores, which is the point. If you sell on Shopify or Wix, you already have an llms.txt whether you thought about it or not.

Here is the honest catch, and it is exactly why this matters for you: WordPress and WooCommerce do not generate one for you. There is no native llms.txt in WordPress core or in WooCommerce. So a WooCommerce store, which might have a far better catalog than a random Shopify shop, is currently the one showing up to AI tools as raw, cluttered HTML while its Shopify competitors hand over a clean, structured file automatically. That gap is the reason a plugin exists.

So the two facts sit together honestly. Reading adoption across the whole web is still low, and the platforms your competitors use are already shipping this by default. If your read of the trend is that machine-readable stores become normal, you want a clean one published now, not a scramble later once it is expected.

Why publish an llms.txt anyway?

Three reasons hold up.

It puts you ahead of an emerging standard. The companies you would benchmark yourself against are already publishing. The sites that already speak the format win the moment adoption broadens, and you are not left retrofitting your whole catalog under pressure. Publishing now is a position, not a chore.

It is genuinely useful to agents and tooling right now. If any part of your audience uses AI shopping agents, research agents, or coding tools that fetch content, a clean llms.txt helps them today. For online stores, content businesses, documentation-heavy sites, and developer products alike, this is a present-tense benefit, not a bet on the future.

It gives you a clean, controlled representation of your site. Rather than letting a tool guess at your content from cluttered HTML, full of navigation, cookie banners, and layout markup, you hand it an accurate, curated version and decide what matters. That control is the whole point, and it is worth getting right rather than approximating by hand.

What you should not do is publish one expecting a ranking boost or guaranteed citations. That promise is not backed by evidence today, and anyone who tells you otherwise is selling ahead of the facts. The honest value is real, but it is about being machine-readable and being early, not about gaming search.

How to create an llms.txt file on WordPress

You have two paths.

By hand, you write the Markdown yourself and upload the file to your site’s root so it is served at yourdomain.com/llms.txt. This is fine for a small, static site, and it is a good way to learn the format. The catch is maintenance: every time your content changes, the file drifts out of date, and a stale llms.txt misrepresents your site.

Automatically, a plugin reads your existing pages, posts, and products, converts them to clean Markdown, and writes both llms.txt and llms-full.txt for you. On WordPress that is the far more practical route, especially once you have more than a handful of pages or a store to describe. Our own Project LLMs.txt plugin does exactly this: install it, click Generate, and your files are live in a couple of minutes, WooCommerce products included.

Frequently asked questions

Is llms.txt an official web standard?

No. It is a community convention, proposed by Jeremy Howard of Answer.AI in 2024 and documented at llmstxt.org. No standards body has adopted it, and no AI company is obliged to read it.

Does llms.txt improve my Google rankings?

No. Google has stated that the file has no effect on Search rankings or on AI Overviews, in either direction. Publish it for the tools that do read it, not for Google.

Do I need both llms.txt and llms-full.txt?

Not strictly, but publishing both covers more situations. The index is small and cheap for a tool to fetch, while the full file spares it from making dozens of follow-up requests to read your actual content.

Does WordPress create an llms.txt automatically?

No. Neither WordPress core nor WooCommerce generates one. Shopify does, and Wix does on its premium ecommerce plans, which is why WordPress stores are currently the ones handing AI tools raw HTML while their competitors hand over clean Markdown.

Where does the llms.txt file go?

At the root of your domain, served at yourdomain.com/llms.txt, the same place robots.txt lives. Tools look there and nowhere else, so a copy in a subdirectory will not be found.

Will llms.txt conflict with my robots.txt or sitemap.xml?

No. The three files do different jobs and none of them overrides another. robots.txt controls crawler access, sitemap.xml lists your URLs for search engines, and llms.txt describes your content for AI tools.

How often should I update it?

Whenever your content changes in a way that would make the file wrong. A stale llms.txt actively misrepresents your site, which is arguably worse than not having one, so the file is only as valuable as it is current.

llms.txt in summary

llms.txt is a simple Markdown file that describes your site cleanly for AI tools. It is used today by AI shopping and research agents, coding assistants, and other AI tooling, and it is now generated automatically by platforms like Shopify and Wix. It is not used by Google Search, and there is no evidence it improves your visibility in consumer AI answers. Adoption on the reading side is still early.

Publish one because it helps the tools that already read it, because it gives you a clean and controlled version of your site, and because the platforms your competitors use are already doing it for them. Do not publish one expecting it to change your Google rankings. Within those honest bounds, it is worth doing, and worth doing properly.

If you are still sorting out how this file relates to the ones you already have, see llms.txt vs robots.txt vs sitemap.xml. For an honest look at how much it is actually being read, see the real state of llms.txt adoption in 2026.

Project LLMs.txt is a WordPress plugin that generates your llms.txt and llms-full.txt automatically, WooCommerce included. It is free to install and takes about two minutes to set up.