llms.txt: the practical guide

Last updated: September 2026

The format

Three conventions cover nearly every good llms.txt in the wild:

1. An H1 with the site or product name.
2. A blockquote (lines starting with >) with a one-paragraph factual summary — what it is, who it is for, key facts like pricing.
3. H2 sections of markdown links, each with a one-line description of the target page.

Optionally, publish a fuller llms-full.txt alongside it and link the two.

Worked example

# Acme Invoicing

> Acme Invoicing is online invoicing software for freelancers. Plans: Starter $12/mo, Pro $29/mo.

## Product

- [Features](https://acme.example/features): recurring invoices, late-fee automation, 40+ currencies.
- [Pricing](https://acme.example/pricing): full plan comparison.

## Resources

- [Late-fee guide](https://acme.example/guides/late-fees): how freelancers charge late fees in 12 countries.
- [API docs](https://acme.example/docs): REST API reference.

How to publish it

1. Write the file (keep it factual — it is a map for models, not ad copy).
2. Serve it at /llms.txt with Content-Type: text/markdown (or text/plain; both work).
3. Verify it is reachable in an incognito window and over HTTPS.
4. Run our free audit — it checks llms.txt presence as its first check.

Honest take: does it matter?

llms.txt is not an official standard — no W3C, no IETF, and no AI lab has formally committed to reading it. Anyone telling you it is a "ranking factor" is overselling. But the cost is ~20 minutes and the file is already checked by AI-readiness tooling and several crawlers. Think of it as part of the standard checklist — llms.txt, open AI-crawler rules, structured data, quotable content — not a magic fix on its own. Our own llms.txt is public if you want a real-world example.

Frequently asked questions

llms.txt vs robots.txt — do I need both?

Yes, they do different jobs. robots.txt controls whether crawlers (including GPTBot, ClaudeBot, PerplexityBot) may fetch your pages. llms.txt controls what clean summary they find when they do.

Will llms.txt leak anything sensitive?

Only what you write in it. It is a public file — never put internal URLs, API keys, or non-public information in it.

How often should I update it?

Whenever your key pages, pricing, or positioning change. A stale llms.txt is worse than none, because models may quote outdated facts with confidence.