llms.txt

Tell models what your site is, in your own words.

One markdown file at the root of your domain, describing your site in the words you would use yourself — instead of leaving a language model to reconstruct it from your navigation and your footer.

Before you start

It is a proposal, not a standard.

llms.txt was proposed in 2024 and no standards body has ratified it. But it is being used: crawlers request the file, tooling reads it, and providers ship one themselves — Anthropic publishes an llms.txt for its own documentation. "Nobody reads it" is simply wrong.

What is genuinely open is the effect. No provider documents how much weight the file carries when sources are picked, so nobody can put a number on what it earns you. It costs twenty minutes, it breaks nothing, and writing it forces you to state plainly what your product does and what it costs. That is why it carries moderate weight in our score today rather than none — and why that weight goes up as adoption does.

The format

What goes in it.

Plain markdown at https://your-domain.com/llms.txt. One H1 with your name, a blockquote summarising what you do, then sections with links. No fixed schema — these are the parts that consistently make sense:

  • A one-line summary a model can quote verbatim
  • What you sell, and what it costs — in numbers, not adjectives
  • Links to the pages that matter, with a word on what each contains
  • What you explicitly do not do — it prevents wrong assumptions
/llms.txt
# My Company

> Cloud inventory management for small retailers. Items, documents,
> stock and invoices in one system. Operated in the EU.

## Product

- **Basic** — €29 per month. Item master, documents, one warehouse.
- **Pro** — €79 per month. Multiple warehouses, multi-currency, API.

## Pages

- [Home](https://your-domain.com/)
- [Pricing](https://your-domain.com/pricing) — all plans compared
- [Contact](https://your-domain.com/contact)

## What we do not do

- No payroll, no bookkeeping.

## Contact

hello@your-domain.com
Common mistakes

What we find when we check it.

Save yourself the blank page

We draft it from your own site.

After a scan we already know your routes, your titles and your page inventory. Trust customers get a generated llms.txt draft from exactly that — review it, adjust the wording, upload it. Every later scan then checks whether it still matches the site.

What Web Ready checks

Questions

Where does the file go?
At the root of your domain, reachable as /llms.txt, served as plain text or markdown.
Is llms-full.txt a separate thing?
It is an optional longer variant holding full page content rather than links. Useful for documentation sites, unnecessary for most marketing sites.
Does it replace robots.txt?
No, and it cannot. robots.txt controls access, llms.txt describes content. A model that is blocked in robots.txt will not read your llms.txt either.
Does anyone actually read it?
Yes. Crawlers request the file, tooling consumes it, and Anthropic among others publishes one for its own docs. What nobody documents is how heavily it weighs when sources get picked — that part we do not pretend to know.
Get it

Web Ready comes with Trust.

One tier, one price. There is no separate Web Ready subscription to work out.

Trust
€24 every 4 weeks
  • Public security badge, auto-updating
  • Daily re-scan of your site
  • Web Ready report — SEO and AEO, private to you
  • AI-fix prompts — paste into Claude or GPT
  • Trigger a scan on demand (rate-limited)
Subscribe to Trust — €24 Compare all tiers
✓ EU-hosted · GDPR ✓ Cancel anytime ✓ No credit-card trial trap

Price excl. VAT — added at checkout where applicable. B2B only.