What Is llms.txt? A Plain-English Guide for 2026

If you work in SEO, you have probably seen llms.txt come up a lot lately, usually with a lot of confidence and not much agreement. Here is the plain version: what it is, what goes in it, whether it actually does anything, and how to write one that is worth publishing.

What is llms.txt?

llms.txt is a plain Markdown file you put at the root of your site, at yourdomain.com/llms.txt. The idea, proposed at llmstxt.org, is to give large language models a curated guide to your site: a short summary plus a hand-picked set of links to the pages that actually matter, in a format that is cheap for a model to read.

Think of it as a counterpart to robots.txt and sitemap.xml, but aimed at LLMs instead of search crawlers. robots.txt says what bots may access. sitemap.xml lists everything. llms.txt is meant to be the opinionated short list: here is who we are, and here are the pages worth reading.

Some sites also publish a larger llms-full.txt, which inlines the actual content rather than just linking to it.

What goes in it

The format is deliberately simple. Reading the spec closely, only one thing is strictly required:

• An H1 with the site or project name. This is the only required element.
• A blockquote summary (recommended): one line capturing what the site is.
• Sections under ## headers, each a list of [name](url) links, optionally with a short note after each link.
• An Optional section (a literal ## Optional header) for links that can be skipped if a model is short on context.

A minimal, well-formed file looks like this:

# Acme Analytics

> Acme Analytics is a privacy-first web analytics tool for small teams.

## Docs

- [Getting started](https://acme.com/docs/start): install and first dashboard
- [API reference](https://acme.com/docs/api): endpoints and auth

## Optional

- [Changelog](https://acme.com/changelog)

That is the whole idea. A title, a summary, and a curated set of links with context.

Does Google use llms.txt?

No, and this is the part worth being clear about. In June 2026, Google clarified its guidance: Google Search does not use llms.txt, and publishing one will neither help nor hurt your rankings. Their own words are that it is “fine if you want to maintain these files for other services or systems that use them,” but it is not a Search ranking signal.

So if anyone sells you llms.txt as a Google ranking hack, that is not accurate. It is not one.

Who actually reads it, then?

The audience is the set of AI tools and agents that have chosen to support the format: certain assistants, agentic browsers, and dev tools that pull a site’s llms.txt to orient themselves before reading further. It is opt-in on the consumer side, and adoption among the big platforms is still uneven.

One concrete place it shows up: Chrome’s Lighthouse now includes an agentic-browsing audit that checks for a valid llms.txt. So it has started to appear in mainstream tooling, even though Search itself ignores it.

The honest summary: this is an emerging, optional standard with real but still-limited readership. Publishing one is a low-cost bet on where AI-driven discovery is heading, not a guaranteed payoff today.

Should you create one?

If it is cheap for you to do well, yes. A clean llms.txt gives the systems that do read it an accurate map of your site, and it costs you very little. Just go in with the right expectations: you are doing it for AI tools and agents, not for Google rankings.

Two things matter more than having one at all:

1. Make it a real one. A curated title, summary, and a handful of meaningful sections. A file that just dumps every page on your site is not the guide the format is for.
2. Make your signals agree. Do not publish an llms.txt inviting AI tools in while your robots.txt blocks the AI crawlers. That contradiction is more common than you would think.

I built two free tools for this: an llms.txt generator to write a clean, spec-correct file, and an AI crawler access checker to see which AI bots your robots.txt currently allows or blocks.

What most people get wrong

I crawled the top 10,000 sites to see what adoption actually looks like, and the findings are a useful checklist of what to avoid. Nearly half of the files that exist are missing the basic structure, a meaningful share are auto-generated dumps from an SEO plugin, and a real slice of sites publish an llms.txt while simultaneously blocking AI crawlers in robots.txt.

If you want the full picture, the data is here: The State of llms.txt.

The short version: having an llms.txt is easy. Having a good one, and keeping it consistent with the rest of your site, is what actually makes it worth publishing.