Reference · Readable

llms.txt

A one-file map that tells an agent which of your pages are worth reading.

What it is

/llms.txt is an emerging convention: a markdown file at a predictable path with your name, a short summary, and a curated list of links with a sentence of context each.

It is not robots.txt and it is not a sitemap. robots.txt says what is permitted; a sitemap lists everything; llms.txt says what actually matters.

Why an agent cares

An agent lands on your site with a limited context budget. Left to crawl, it spends that budget on your navigation, your cookie banner and your careers page.

Be clear about what it does not do. It is not a ranking or citation lever: an Ahrefs study of 137,210 domains found 97% of llms.txt files received no requests at all in May 2026, and most of the requests that did arrive came from SEO audit tools and ordinary crawlers rather than AI assistants. It is an afternoon of writing, and it is no longer rare: most of the well-known SaaS sites in our benchmark already publish one.

How to fix it

Publish /llms.txt as plain markdown. Curate ruthlessly: fifteen good links beat two hundred.

# Acme

> Payroll for UK companies under 50 people. Runs payroll,
> files with HMRC, and handles pensions automatically.

## Docs
- [Quickstart](https://acme.com/docs/quickstart): Run your first payroll in ten minutes.
- [API reference](https://acme.com/docs/api): REST endpoints, auth, rate limits.

## Company
- [Pricing](https://acme.com/pricing): Per-employee monthly pricing, no setup fee.
- [Security](https://acme.com/security): SOC 2 Type II, data residency, sub-processors.

How we test it

We fetch /llms.txt and /llms-full.txt and require real content — a soft-404 that returns your HTML app shell doesn't count.

Related

SitemapAgent manifestAI crawler access

Does your site pass this one?

Run the full scan and find out — along with the other 18.