Implementation Guide

How to Structure Your llms.txt File

The spec is simple. The strategy isn't. This guide covers the exact format, how to organize sections for your site type, what to include, what to skip, and the mistakes that make your file useless.

Last updated: March 2026

The Spec Format

An llms.txt file is plain markdown. Four elements. That's it.

llms.txt
# Your Site Name                    ← H1 title (required)

> A one-line description of your site.   ← Blockquote summary

## Core Resources                       ← H2 section heading

- [About](https://you.com/about): Who we are   ← Link with description
- [Products](https://you.com/products): What we sell

## Optional

- [Blog](https://you.com/blog): Industry insights
- [Careers](https://you.com/careers): Open positions

H1 Title

Required. Use # followed by your site or company name. This is the identity line -- AI uses it to know who you are.

Blockquote Summary

Use > for a 1-2 sentence description. Be specific: what you do, where, and what makes you different. This is the first thing AI reads.

Section Headings

Use ## to group links by topic. The first section should be your most important pages. AI systems respect this hierarchy.

Descriptive Links

Format: - [Name](URL): Description. The description after the colon is critical. It tells AI what the page is about, not just what it's called.

Section Strategy by Site Type

The sections you use depend on what kind of site you run. We've analyzed 840+ real implementations and identified the patterns that work best for each type.

SaaS & Software

AI gets asked about tools constantly. "What's the best project management software?" "How does Stripe handle subscriptions?" Your llms.txt needs to lead with product capabilities.

## Products
- [Feature A](URL): What it does and who it's for
- [Feature B](URL): The specific problem it solves

## Documentation
- [Getting Started](URL): Quick-start integration guide
- [API Reference](URL): Complete API documentation

## Resources
- [Case Studies](URL): How customers use the product
- [Blog](URL): Product updates and industry analysis

## Company
- [About](URL): Team, mission, founding story
- [Pricing](URL): Plans and feature comparison

E-Commerce

Don't list every product. List categories and buying guides. AI needs to understand what you sell and why someone should buy from you, not every SKU in your inventory.

## Product Categories
- [Running Shoes](URL): Men's and women's running shoes
- [Trail Shoes](URL): Off-road and hiking footwear
- [Sale](URL): Current markdowns and clearance

## Buying Guides
- [How to Choose Running Shoes](URL): Fit guide by foot type
- [Trail vs Road Shoes](URL): Which type you need

## About
- [Our Story](URL): Founded in 2010, based in Portland
- [Shipping & Returns](URL): Free shipping over $75, 30-day returns

## Optional
- [Store Locator](URL): Physical retail locations
- [Size Chart](URL): Measurement guide for all shoe types

Blog & Content Sites

Organize by topic, not by date. AI doesn't care that you published something last Tuesday. It cares about topic authority. Lead with your best work.

## Cornerstone Content
- [Complete Guide to X](URL): The definitive resource on X
- [Y vs Z Comparison](URL): In-depth analysis with benchmarks

## Topic: AI & Machine Learning
- [Understanding LLMs](URL): How large language models work
- [AI in Business](URL): Practical applications for SMBs

## Topic: Web Development
- [Frontend Frameworks](URL): React, Svelte, Vue compared
- [Performance Guide](URL): Core Web Vitals optimization

## About
- [About the Author](URL): Background and expertise
- [Newsletter](URL): Weekly insights on technology

Professional Services

Location, credentials, and expertise areas. When AI is asked "best dentist in Austin," it needs to know your specialties, your location, and why you're credible. Spell it out.

## Services
- [Family Dentistry](URL): Cleanings, exams, preventive care
- [Cosmetic Dentistry](URL): Veneers, whitening, smile design
- [Emergency Care](URL): Same-day emergency appointments

## About Us
- [Our Team](URL): 3 dentists, 15+ years combined experience
- [Locations](URL): Two Austin offices -- North and South

## Patient Resources
- [Insurance](URL): Accepted providers and payment options
- [New Patients](URL): First visit information and forms

## Optional
- [Reviews](URL): Patient testimonials and ratings
- [Blog](URL): Oral health tips and practice news

Portfolio & Freelancer

Lead with what you do and your best work. AI gets asked "who does X in Y city" all the time. Your file needs to answer that question in under 10 seconds.

## Work
- [Project Alpha](URL): Branding and web design for fintech startup
- [Project Beta](URL): E-commerce redesign, 40% conversion lift
- [Project Gamma](URL): Mobile app UI for healthcare platform

## Services
- [Brand Design](URL): Logos, identity systems, brand guidelines
- [Web Design](URL): Responsive websites and web applications

## About
- [About Me](URL): 8 years in design, based in Brooklyn
- [Contact](URL): Availability and booking information

What to Include vs. What to Skip

Curation is the game. An llms.txt with 200 links is worse than one with 25 well-chosen links. AI systems process information better when the signal-to-noise ratio is high.

Include

  • + Homepage and about page
  • + Core product or service pages
  • + Category pages (not every product)
  • + Cornerstone content and guides
  • + Pricing and plans
  • + Contact and location info
  • + Case studies and social proof
  • + Key policies (shipping, returns)

Skip

  • - Login and account pages
  • - Admin or dashboard URLs
  • - Duplicate or paginated content
  • - Individual product pages (use categories)
  • - Tag and archive pages
  • - Utility pages (search, 404, etc.)
  • - Outdated or thin content
  • - Legal boilerplate (unless service-relevant)

Writing Descriptions That Work

The description after each link is doing more work than you think. It's the context AI uses to decide whether to follow that link and how to represent that page. A good description tells AI what the page is about and why it matters. A bad one wastes the opportunity.

Bad - [Products](https://example.com/products): Our products

Tells AI nothing it didn't already know from the link text.

Good - [Running Shoes](https://example.com/shoes): Men's and women's road and trail running shoes, sizes 5-15

Specific. AI now knows the product type, audience, and range.

Bad - [Blog](https://example.com/blog): Check out our blog

Marketing speak. Zero information for AI.

Good - [Blog](https://example.com/blog): Weekly articles on JavaScript performance, React patterns, and web accessibility

Now AI knows the topics and can cite specific articles for relevant queries.

Our AI Enhancement feature rewrites your descriptions to maximize clarity and citation potential. It reads each linked page and generates descriptions optimized for how LLMs process content.

When to Use llms-full.txt

The spec defines two files. llms.txt is the concise version -- your site's executive summary. llms-full.txt is the expanded version with more detail.

Think of it this way: llms.txt is what you'd tell someone in an elevator. llms-full.txt is what you'd tell them over coffee.

When llms-full.txt makes sense

  • -- Documentation-heavy sites (SaaS, developer tools, technical platforms)
  • -- Content sites with 100+ quality pages that merit inclusion
  • -- E-commerce sites where product detail matters for AI recommendations
  • -- When you want to include longer descriptions with more context per page

Most sites should start with just llms.txt. Add llms-full.txt later if you need more depth.

Common Mistakes

We've reviewed hundreds of llms.txt files. These mistakes show up constantly. Check your file against this list and follow our best practices guide for the full picture.

URL Dumping

Listing every URL on your site without sections or descriptions. This is the most common mistake. AI can't distinguish your homepage from your 404 page if you treat them the same.

Missing Summary

Skipping the blockquote summary. This is the single most important line in the file. It's the first thing AI reads to understand what your site is about. Don't skip it.

Vague Descriptions

"Click here to learn more" or "Our products page." These tell AI nothing. Every description should answer: what is on this page and why would someone need it?

Never Updating

Deploying once and forgetting about it. Your site changes. Your llms.txt should too. Stale files point AI to dead links and missing content. That hurts your credibility.

Wrong File Location

Putting the file in a subdirectory or behind authentication. It must be accessible at yourdomain.com/llms.txt or yourdomain.com/.well-known/llms.txt. No auth. No redirects.

Or Just Let Us Handle It

Everything on this page? Our generator does it automatically. Enter your URL, we scan your sitemap, and we build a spec-compliant file with proper sections, descriptions, and structure.

You can review and edit every line. Or just download it and deploy. Takes 30 seconds either way.

Build Your llms.txt Now

Free to generate. Takes 30 seconds. No credit card. Get a spec-compliant file with proper structure, sections, and descriptions -- ready to deploy.