Guides

How to validate your llms.txt

Seven short steps, in order. Most problems are on the first line or in the links.

Check your llms.txt now

Free, no sign-up. We read your /llms.txt the way an assistant would, and show each problem with its line and the fix.

Check your whole site: what an assistant can do with your business, and what's missing.

Seven steps

  1. Open it
    Open yoursite.com/llms.txt in a browser. You should see plain text, not your homepage or an error page.
  2. Check line 1
    Line 1 is your name after one #: # Your Business. It's the one line the format requires.
  3. Add a summary
    Below it, one line that starts with >: what you do, and where. Optional, but readers like it.
  4. Make every link a link
    Under each ## heading, one link per line: - [Name](https://…): a short note. Use full https addresses.
  5. Run a validator
    Put your website into the llms.txt checker at the top of this page, and fix each line it names.
  6. See what Lighthouse says
    Optional: in Chrome, open DevTools, then Lighthouse, tick the Agentic Browsing category and run it.
  7. Check again
    Upload the fixed file, then check again. A cache may show the old file for a little while.

Send these steps to my web person

Common errors and how to fix them

What you seeThe fix
Your homepage, or a “page not found” page, at /llms.txt There's no file yet. Upload a file named llms.txt to your site's root, or turn it on in your platform.
Line 1 isn't # Your name Make line 1 your business's name after one # and a space. Nothing but blank lines may come before it.
- Menu: https://… Write it as a link: - [Menu](https://…): lunch and dinner. Chrome's Lighthouse fails a file with no Markdown link.
### headings, or a heading between your name and your sections Keep one # line for your name, and use ## for every section.
http:// or /menu Use the full address with https: https://yourbusiness.com/menu.
A link leads to a missing page Fix the address, or remove the line.
Your robots.txt blocks /llms.txt Allow it: readers that obey robots.txt won't fetch a blocked file. The AI crawler checker shows the line.
Error 403 or 429, or a “checking your browser” page Your host or firewall stops readers. Ask it to let /llms.txt through.
Odd characters, such as é instead of é Save the file as UTF-8 plain text.

What a valid file looks like

Short, with your name first and one link per line:

# Crumb & Co. Bakery

> Bakery and café for sourdough, pastries and celebration cakes. Open every day from 7:00.

## Visit and order

- [Menu and prices](https://crumbandco.example/menu): breads, pastries and coffee
- [Order a cake](https://crumbandco.example/cakes): three days' notice
- [Opening hours and how to find us](https://crumbandco.example/visit)

## Optional

- [Our story](https://crumbandco.example/about)

More examples: a restaurant, a salon, a clinic, a shop, a trade and more.

What a valid file does, and doesn't

Google Search doesn't use llms.txt. Chrome's Lighthouse audits it, and some AI tools read it.

A valid file helps an assistant read about you. To let it ask, book or order, you need a door: a tool, an API or an agent inbox that takes the request. More in what llms.txt does, and doesn't.

Questions

What makes an llms.txt valid?

Per llmstxt.org, only one thing is required: a # heading with your name. A good file adds a > summary and ## sections of [name](url) links.

Does Google check my llms.txt?

Google Search doesn't use it. Chrome's Lighthouse (13.3 and later) audits it under Agentic Browsing: a # heading, at least 50 characters and one Markdown link.

Is llms-full.txt required?

No. llms-full.txt is a community convention, not part of the llms.txt proposal.

Can my llms.txt live in a folder?

Yes. Since version 2 of the proposal (August 2026), a file at a path such as /docs/llms.txt covers the pages under that path.