← Back to Blog
Analytics & SEO5 min read

What Is llms.txt? Why It Matters, and How to Set One Up

Learn what llms.txt is, why Google added an Agentic Browsing score in PageSpeed Insights, and follow a step-by-step guide to setting up an llms.txt file to get cited by ChatGPT, Perplexity & Gemini.

Naveen Gaur
Naveen Gaur
August 11, 2026

If you've run a PageSpeed Insights report recently, you may have noticed a new score sitting next to Performance, Accessibility, Best Practices, and SEO: Agentic Browsing. Unlike the others, it's not a number out of 100 — it's shown as a ratio, like 2/3 or 3/3. And one of the fastest ways to move that ratio up is a small text file called llms.txt.

Here's what it is, why Google added a score for it, and how to set one up on your own website.


What llms.txt Actually Is

llms.txt is a plain text file that sits at your domain's root — yoursite.com/llms.txt — the same way robots.txt does. Instead of talking to traditional search engine crawlers, it is written specifically for Large Language Model (LLM) systems: AI chatbots, conversational search engines, and AI browsing agents that fetch and read your site on someone's behalf.

It serves as a short, structured summary of what your business does, what services or products it offers, and which pages matter most — formatted so an AI model can quickly understand your website without having to piece it together from complex navigation menus, client-side JavaScript, or human-targeted marketing copy.


Why Google Added This to PageSpeed Insights

Agentic Browsing audits were introduced to Lighthouse in May 2026, and Google PageSpeed Insights incorporated them shortly after. The reasoning is straightforward: a rapidly growing share of web traffic today isn't a person clicking around search results — it's an AI agent browsing on their behalf to research, compare providers, or complete tasks. Adobe Analytics reported AI-driven referral traffic to online businesses grew by 393% year-over-year in Q1 2026 alone.

Google's Agentic Browsing score evaluates three core criteria:

  1. Accessibility Tree Validity: Whether screen readers and AI parsing agents can read your DOM elements cleanly.
  2. Layout Stability: Low Cumulative Layout Shift (CLS), ensuring elements don't move unpredictably while agents attempt interactions.
  3. Presence of llms.txt: Whether an llms.txt file exists at your domain root and passes structural validation checks.

A missing or unreadable llms.txt file is the single most common reason a website scores 2/3 instead of 3/3 on Google's Agentic Browsing checks. For a deeper breakdown of these AI audits, see our guide on Google's PageSpeed Agentic Browsing Score.


Why It's Worth Doing (Even Beyond PageSpeed Scores)

Setting the PageSpeed metric aside, there is a fundamental business reason to implement this file: AI assistants are increasingly the first place potential customers ask questions"who does X near me," "what is the best option for Y," or "recommend a specialist for Z."

When conversational AI models like ChatGPT, Perplexity, or Gemini answer those queries, they rely on structured context they can crawl and synthesize quickly. A clear, accurate llms.txt file ensures your business is:

  • Accurately summarized rather than hallucinated or misconstrued.
  • Directly cited when recommendations are displayed.
  • Included in category queries rather than skipped in favor of competitors whose sites are easier for AI to digest.

If your site is also experiencing traditional search ranking issues, check our guide on why your WordPress site is not ranking on Google.


Step-by-Step: How to Create and Publish llms.txt

1. Write a Factual Business Summary

Skip corporate taglines and buzzwords. State plainly what the business does, target audience, credentials, service areas, and pricing models so an AI model can quote these details directly.

2. List Your Primary Pages with One-Line Descriptions

Pull key URLs from your navigation menu — Home, Services, Pricing, About, Contact. Provide full canonical HTTPS URLs for each.

3. State Service Boundaries and Contact Info

Specify whether you operate locally, nationally, or remotely. Ambiguity forces AI models to guess — and guessing often leads to incorrect recommendations.

4. Format Links in Standard Markdown

The audit specifically checks for Markdown link syntax: [Page Name](https://example.com/page), not plain-text URLs (Page Name: https://example.com/page). Plain URLs will fail validation checks.

5. Upload to Your Domain Root

llms.txt must be placed directly at yoursite.com/llms.txt, not inside plugin directories or subfolders. On WordPress, this requires uploading the file via cPanel File Manager, hPanel, or SFTP into your public_html directory.

6. Verify Live Access

Open yoursite.com/llms.txt directly in your browser. Ensure it returns HTTP 200 with raw text, rather than a 404 error page or a redirected homepage.


A Ready-to-Use llms.txt Template

# Business Name

> Factual one-line summary of what the business does and who it serves.

Two to three sentences of specific, checkable details — services offered, credentials, location/service area, and booking process.

## Main Pages
- [Home](https://example.com/)
- [Services](https://example.com/services/)
- [About Us](https://example.com/about/)
- [Contact](https://example.com/contact/)

## Contact & Location
- Location: City, State (or Fully Remote)
- Email: hello@example.com
- [Book an Appointment](https://example.com/booking)

Don't Want to Handle Technical Setup Yourself?

If you prefer to avoid editing root server directories, configuring SFTP, or debugging validation errors, we offer a complete done-for-you service.

We write, format, upload, and verify your website's llms.txt file live on your site within 48 hours for $10.

👉 Get Your Custom llms.txt Setup Service ($10)



Related Reading: Google's New PageSpeed Agentic Browsing Score Explained | Why Your WordPress Site Is Not Ranking on Google | WordPress PageSpeed Optimization Case Study

Leave a Comment

Comments are moderated before appearing on the site.

Full-Stack & Cloud Architecture

Need custom full-stack development or cloud architecture help?

I build high-performance web applications, custom APIs, and scalable cloud infrastructure for tech founders. Fast turnaround, direct access, no agency overhead.

Direct access to Naveen Gaur • Response within 24 hours