What is llms.txt?

The short definition

llms.txt is a proposed standard: a single Markdown file served at /llms.txt that gives large language models a clean, curated summary of your site. Instead of forcing a model to crawl and guess its way through your marketing pages, you hand it a concise map — what you do, who you’re for, and the URLs that matter.

Why it exists

Web pages are noisy. Navigation, scripts, cookie banners, and marketing flourishes make it hard for a model to extract what your product actually is. llms.txt solves this the way robots.txt once solved crawler guidance: a predictable, machine-first file in a known location, written in plain language a model can parse without ambiguity.

What goes in it

A good llms.txt is short and structured:

• An H1 with your name and a one-line summary (often as a blockquote) of what you do.
• Short sections — what you do, what you optimize, pricing, contact — in clear prose.
• Curated links to the pages that matter most, each with a short note on what it covers.

The goal isn’t to dump your whole site — it’s to give the model the smallest, clearest version of the truth. (Nivonto’s own file lives at /llms.txt — we dogfood this.)

Does llms.txt actually help?

Adoption is still early and not every engine reads it yet — so treat it as one signal, not a silver bullet. It costs little to add, removes ambiguity for the models that do use it, and signals that you’ve thought about being machine-readable. It works best alongside the rest of the AEO toolkit: structured data, buyer-question content, and consistent presence across the web.

llms.txt vs robots.txt

They solve different problems. robots.txt tells crawlers where they may go; llms.txt tells models what your site means and where the important content is. For AEO you want both — a robots.txt that welcomes AI crawlers, and an llms.txt that explains you clearly once they arrive.