Skip to content

AI SEO

llms.txt: what it is, what it does, and what it does not do yet

llms.txt is a proposed plain-text file that gives language models a curated map of your site. Adoption is unconfirmed, the cost is near zero, and the structural discipline it forces is worth more than the file itself.

Mawkaii 6 min read

Key takeaway

llms.txt is a proposed Markdown file at your domain root that lists your most important pages with short descriptions, giving language models a clean map of your site. No major model provider has confirmed it as a ranking or retrieval signal. It costs very little to publish, it forces you to decide what your most important content actually is, and it positions you correctly if adoption grows — but it should be treated as a cheap bet, not a proven lever.

Every few months the SEO industry finds a new file to be excited about. This is the current one. It deserves a measured answer rather than either the breathless or the dismissive version.

What it is

llms.txt is a proposed convention: a Markdown file at the root of your domain that gives language models a curated, human-readable map of your most important content.

Where robots.txt says what may be crawled and sitemap.xml lists every URL mechanically, llms.txt is editorial. You choose what matters and describe it in a sentence.

A minimal example:

# Mawkaii

> A design and development studio building fast, findable digital
> products — and the SEO and AI SEO programmes that bring people to them.

## Services
- [Design & Development](https://mawkaii.com/services/design-and-development/): Web and mobile products, designed and built by one team.
- [AI SEO](https://mawkaii.com/services/ai-seo/): Getting cited inside ChatGPT, Perplexity and AI Overviews.

## Insights
- [What is AI SEO?](https://mawkaii.com/blog/what-is-ai-seo/): A working definition and what it changes.

What it does not do, yet

No major model provider has publicly confirmed that llms.txt affects retrieval or citation. Anyone telling you otherwise is extrapolating.

Some crawlers fetch it. That is not the same as weighting it.

So why publish one? Three honest reasons:

  1. The cost is close to zero. Ours is generated automatically at build time from the same content that produces the sitemap. It costs nothing to maintain.
  2. It forces a useful decision. Writing one means deciding which twenty pages actually represent your business. Most teams have never done that explicitly, and the exercise usually surfaces content that should not exist.
  3. The option value is asymmetric. If adoption grows, you are already there. If it does not, you spent an hour.

That is the whole case. Anything more emphatic is overselling.

How to write a good one

  • Lead with a one-paragraph description of what your organisation actually does, in plain language, under a blockquote.
  • Group by intent, not by your internal site structure. Services, evidence, insights.
  • Describe each link in one sentence that would make sense read aloud with no other context.
  • Be ruthless. Twenty to forty links. If you list two hundred you have rebuilt your sitemap and thrown away the only advantage the format has.
  • Generate it from your content source, not by hand, or it will be stale within a quarter.

Ours

You can read our llms.txt. It is generated at build time from the same content collections that render the site, so it cannot drift out of date.

That is the real point of the file, incidentally: not the bet on adoption, but the discipline of having a single place where your site explains itself clearly enough that a stranger — or a model — could summarise your business correctly.

  • llms.txt
  • AI SEO
  • technical SEO

We do this for a living.

If this is the sort of thinking you want on your side, the first conversation is free.