Does llms.txt actually do anything?

The short answer

Not much yet. No major AI platform has confirmed it reads llms.txt, and none of them treat it the way search engines treat robots.txt or sitemaps. It costs ten minutes to publish and does no harm, but it will not get you cited. Fix your robots file and your entity data first.

What it claims to do

The proposal is a markdown file at your root that gives models a clean, curated summary of your site and its most important pages. The intent is sensible: give a model a shortcut to the good stuff instead of making it parse your navigation.

The problem is adoption. A standard only works when the consumers of it agree to consume it, and so far they have not publicly committed.

Why publish one anyway

It is cheap, it is a useful forcing function for writing down what your site is actually about, and if adoption comes you are already there. Treat it as a ten minute job rather than a strategy.

The genuinely useful side effect is that writing it makes you articulate your entity clearly, which you then reuse in your schema and your about page.

What to do instead

Spend the effort on things that demonstrably work now. Allow the AI crawlers in robots.txt. Publish complete Organisation schema with sameAs links. Structure pages answer-first. Earn coverage on sites the models already read. Those four move the needle. A markdown file at your root does not, yet.

Quick answers

Will llms.txt ever matter?

Possibly. If one major platform adopts it, the rest tend to follow. Having one in place costs nothing and means you are not scrambling later.

Where does it go?

At the root of your domain, at /llms.txt, in plain markdown. Keep it short and link to your genuinely important pages rather than everything.

More on AI Search

Start here

Ready for
some fame?

Tell us the brand and the number. You get a straight answer on whether we can hit it.

hello@blamethefame.com