llms.txt generator

Website address Generate

What this generates, and what it does not

It reads one page — the one you give it — and builds the standard shape from real material: an h1 with the site's own name, a blockquote from its own description, and its own links grouped by the first segment of each path, because that is how nearly every site is already organised. Policy pages are put in their own group at the end.

It leaves the annotations blank on purpose. The one-sentence description after each link is the part of the format that does the work, and only you know what is on those pages. A generator that filled them in would be inventing facts about a site it fetched once, and the file would be believed. This is the same reason our own llms-txt-quality check scores a link-free file at half marks rather than passing it.

The format, in full

One h1, a blockquote, then sections of annotated links. That is the whole specification.

# Acme Robotics

> Acme builds warehouse robots for companies that would rather not buy a warehouse.
> Founded 2019, based in Rotterdam, shipping to the EU and the UK.

## Product
- [How it works](https://acme.test/product): The arm, the rail, and what fits on them.
- [Pricing](https://acme.test/pricing): Three plans, exact prices, no "contact us".

## Docs
- [Quickstart](https://acme.test/docs/start): Running in about ten minutes.

## Policies
- [Returns](https://acme.test/returns): 30 days, we pay the shipping.

Is it worth publishing one?

Honestly: it is a cheap option rather than a channel. No major assistant has committed to reading the file, and anyone promising that publishing one puts you in ChatGPT is selling something. What is true is that it costs an afternoon, several agent frameworks do fetch it, and almost nobody has one.

Of 292 well-known public sites we measured on 2026-09-06, 64 publish an llms.txt. For comparison, 196 publish a sitemap.xml and 238 publish a robots.txt.

The better argument is the exercise rather than the artefact. To write a useful one you have to decide, in a sentence, what your site is for, and then pick the twenty pages that matter out of the four hundred you have. Most sites find something wrong with their own information architecture on the way. The longer version of this argument, with the evidence.

Questions

What is llms.txt?

A Markdown file at the root of your domain that tells an AI assistant what your site is and which pages matter. It sits alongside robots.txt and sitemap.xml and does a different job from both: robots.txt grants or refuses permission, sitemap.xml lists every URL you have, and llms.txt is the curated short version — the twenty links you would send somebody who asked what you do.

Does anything actually read it?

No major assistant has publicly committed to reading llms.txt. OpenAI, Anthropic and Google have not announced support for it, and anybody telling you that publishing one gets you into ChatGPT is selling something. What is true is narrower: several agent frameworks and documentation tools do fetch it, it costs an afternoon, and writing one forces you to decide in a sentence what your site is for.

Where do I put the file?

At https://yourdomain/llms.txt, served as text/plain or text/markdown, not behind a redirect and not behind a login. On most hosts that means dropping it in the public or static directory.

Why has the draft got no descriptions on the links?

Because only you know what is on those pages. The sentence after each link is the part that does the work — it is what lets an assistant choose the right page without fetching all of them — and inventing one would be inventing a fact about your site. The draft leaves them for you to write, and our own llms.txt check goes on reporting them as missing until you do.

Is this free, and what do you do with the URL?

Free, with no account. We fetch the page once to read its title, description and navigation, and we do not store the URL, publish it anywhere, or add it to any index. Nothing about a site appears publicly unless its owner proves they control the domain and asks to be listed.