Doorbell

Check your website's llms.txt

llms.txt is a short text file that tells AI agents which pages on your site matter. Enter your website: the free scan checks whether yours exists and follows the recommendations Google Lighthouse tests for, and runs the rest of Lighthouse's Agentic Browsing audit.

Free scan

Enter a website address. We run Google Lighthouse's Agentic Browsing audit on it and show what passes and what fails. It usually takes 15–40 seconds.

What Agent Files delivers

Every file is written from your own site: only facts it states, with [placeholders] where something is unknown, and install notes for each file. Here is a shortened sample for a made-up dental practice.

See the sample delivery

llms.txt

A short map of the site for AI agents, with links only to pages that exist.

# Peach Dental

> Family dentist in Atlanta, GA: cleanings, fillings, crowns and implants, with evening appointments on Thursdays.

## Pages
- [Services](https://peachdental.example/services): cleanings, fillings, crowns and implants
- [New patients](https://peachdental.example/new-patients): what to bring and the forms to fill in
- [Contact](https://peachdental.example/contact): address, phone and opening hours

## Booking
- [Book an appointment](https://peachdental.example/book): online booking for new and existing patients

robots.txt

The rules the site already had, kept, plus explicit permission for AI crawlers (shortened).

# Kept from the current file
User-agent: *
Disallow: /wp-admin/

# AI crawlers, allowed by name
User-agent: GPTBot
Allow: /

User-agent: ClaudeBot
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Google-Extended
Allow: /

Sitemap: https://peachdental.example/sitemap.xml

schema-org.jsonld

What the business is, in a form agents read without guessing. Facts the site does not state stay as placeholders.

{
  "@context": "https://schema.org",
  "@type": "Dentist",
  "name": "Peach Dental",
  "url": "https://peachdental.example/",
  "description": "Family dentist in Atlanta, GA.",
  "telephone": "[PHONE]",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "[STREET ADDRESS]",
    "addressLocality": "Atlanta",
    "addressRegion": "GA"
  },
  "openingHours": "[OPENING HOURS]"
}

INSTALL.md (excerpt)

One section per file: where it goes, how to check it, what to replace.

## llms.txt
Where: the root of your site, so it opens at https://peachdental.example/llms.txt.
On WordPress, upload it with your host's file manager; a page or post will not work.
Check: open that address; you should see the text, not a 404 page.

## schema-org.jsonld
Where: inside a <script type="application/ld+json"> tag in your homepage's <head>.
Replace: [PHONE], [STREET ADDRESS] and [OPENING HOURS] with your real details first.
Check: run your homepage through Google's Rich Results Test.

Depending on what your scan finds missing, a delivery can also include sitemap.xml, an agent catalog (ai-catalog.json) and security.txt. MCP and A2A cards and OAuth metadata only apply if your business runs an API or an agent; the install notes explain what they would take instead of inventing them.

Get my files: $99

Why llms.txt matters

Without an llms.txt, an AI agent has no short map of your site and has to guess which pages are worth reading. Google added the check to Lighthouse's Agentic Browsing audit, so it now sits next to accessibility and layout stability in the same report.

The scan shows whether your llms.txt passes or fails and what is wrong with it, the other Agentic Browsing checks with the elements that fail them, and the discovery files below. It is free and takes under a minute. Writing llms.txt and the other missing files for your business is what Agent Files does.

What agents look for next, also checked free

Besides Google's seven agentic-browsing audits, every scan checks the files agents look for next:

The report shows which are in place, broken or missing. Writing the missing ones for your site is what Agent Files does.

Pricing

Card payments are handled by Stripe. Prices in USD.