Sabuj Kundu 6th May 2026

As AI models increasingly crawl and learn from web content, website owners are looking for ways to control how their data is used. This is where llms.txt comes in — a proposed standard similar to robots.txt, but designed specifically for Large Language Models (LLMs).

What is llms.txt?

Proposed by Jeremy Howard (Answer.AI), llms.txt is a Markdown file placed in the root of your website. llms.txt does not enforce rules like robots.txt. Instead, it provides structured guidance to Large Language Models about which content on your site is most important, reliable, and worth referencing. It provides LLMs with a concise, well-structured summary of your site and direct links to the most important resources.

Think of it as:

  • robots.txt → controls search engine crawlers
  • llms.txt → controls AI model usage

llms.txt vs robots.txt vs sitemap.xml

Feature robots.txt sitemap.xml llms.txt
Purpose Crawler access control Page discovery for search engines Guide LLMs with curated content
Format Plain text directives XML Markdown (human + AI readable)
Main Benefit Block unwanted crawlers Better SEO indexing Better AI answers & citations

Why You Need llms.txt

  • Guide AI models toward your most important content
  • Improve the chances of correct citation and attribution
  • Reduce hallucinations by providing canonical sources
  • Make your content easier and cheaper (token-wise) for LLMs to process
  • Prepare your site for Answer Engine Optimization (AEO)

Important: llms.txt Is Not an Access Control File

Unlike robots.txt, llms.txt does not block or restrict bots. It is a recommendation layer, not a permission system.

  • ❌ It cannot prevent AI from accessing your content
  • ❌ It does not define legal usage rights
  • ❌ It is not enforced by crawlers
  • ✅ It helps AI models find high-quality, canonical sources
  • ✅ It improves how your content is understood and cited

What is llms-full.txt?

While llms.txt should stay concise, llms-full.txt can contain a much more detailed, structured version of your content.

  • llms.txt → Summary + key links (fast to read)
  • llms-full.txt → Expanded content (deep context)

This allows AI systems to choose between a quick overview and a deeper crawl depending on their context limits.

Where to Place llms.txt

The file must be placed at the root of your domain:

https://codeboxr.com/llms.txt

Just like robots.txt, it should be publicly accessible.

Basic Structure of llms.txt

Unlike robots.txt, llms.txt uses Markdown format — not plain text directives. It follows a specific structure designed to be both human-readable and easy for AI models to parse. The file consists of three main parts:

  1. An H1 heading with the site or project name
  2. A blockquote with a short description (the elevator pitch)
  3. Sections with H2 headings containing named links to important pages

Here’s a minimal example:

# Codeboxr

> WordPress plugins, Laravel development services, and SaaS solutions for businesses worldwide.

## Docs


- [Plugin Documentation](https://codeboxr.com/docs/): Full documentation for all Codeboxr plugins

- [Getting Started](https://codeboxr.com/getting-started/): Quick setup guides

## Blog


- [Blog](https://codeboxr.com/blog/): Tutorials, updates, and technical articles

## Optional


- [Changelog](https://codeboxr.com/changelog/): Version history and release notes

Key Structural Rules

  • H1 → Name of your site or product (required, only one)
  • Blockquote → A brief summary of what your site offers (required)
  • H2 sections → Named categories of links (e.g., Docs, Blog, API)
  • Links → Named URLs pointing to key pages, with an optional short description
  • “Optional” section → Content that LLMs can skip if context window is limited

Step-by-Step: How to Write llms.txt

Step 1: Add Your Site Name and Description

Start with an H1 heading and a blockquote that clearly describes what your site is about. This is what the AI reads first.

# Codeboxr

> We build WordPress plugins and Laravel-based web applications for businesses across the globe.

Step 2: Add Your Most Important Pages

Group your key pages under H2 sections. Prioritize documentation, product pages, and any content you want AI models to cite or reference correctly.

## Products


- [CBX Petition for WordPress](https://codeboxr.com/product/cbx-petition-for-wordpress/): Petition and signature collection plugin for WordPress

- [CBX HRM](https://codeboxr.com/product/cbxhrm/): Human Resource Management plugin for WordPress

## Docs


- [Plugin Documentation](https://codeboxr.com/docs/): Complete guides and API references for all plugins

## Services


- [Laravel Development](https://codeboxr.com/laravel-development/): Custom Laravel application development for Western businesses

Step 3: Add an Optional Section for Low-Priority Content

The Optional heading is a spec-defined signal. LLMs with limited context windows are advised to skip content under this section if needed.

## Optional


- [Changelog](https://codeboxr.com/changelog/): Product update history

- [Affiliates](https://codeboxr.com/affiliates/): Partner and affiliate program details

Step 4: Link to Markdown Versions of Key Pages (If Available)

If you have .md versions of your documentation or blog posts, link to those instead of or alongside the HTML versions. Markdown is significantly more token-efficient for LLMs to process.

## Docs


- [Getting Started (Markdown)](https://codeboxr.com/docs/getting-started.md): Setup guide in markdown format

Step 5: Keep It Under 10KB

The spec recommends keeping llms.txt concise. If you have extensive content, summarize it here and link out. The goal is a quick, authoritative map — not a full content dump. Use llms-full.txt for the latter.

Complete Real-World Example

Here’s a production-ready llms.txt for a WordPress plugin business:

# Codeboxr

> Codeboxr builds WordPress plugins and custom Laravel web applications for businesses worldwide. Our plugins cover HRM, petitions, booking, job boards, email SMTP, and more.

## Products


- [CBX Petition for WordPress](https://codeboxr.com/product/cbx-petition-for-wordpress/): Create and manage online petitions and signature campaigns

- [CBX HRM](https://codeboxr.com/product/cbxhrm/): Full-featured human resource management plugin for WordPress

- [CBX Booking](https://codeboxr.com/product/cbx-booking/): Appointment and booking management for WordPress

## Services


- [Laravel Development](https://codeboxr.com/laravel-development/): Scalable custom web application development

- [WordPress Plugin Development](https://codeboxr.com/wordpress-plugin-development/): Bespoke plugin development for any use case

## Docs


- [Documentation Home](https://codeboxr.com/docs/): Guides, references, and tutorials for all products

## Blog


- [Blog](https://codeboxr.com/blog/): Technical articles, plugin updates, and development insights

## Optional


- [Changelog](https://codeboxr.com/changelog/): Version history and release notes

- [Contact](https://codeboxr.com/contact-us/): Get in touch with the Codeboxr team

Advanced Examples

Minimal Single-Product Site

# MyPlugin

> A lightweight WordPress plugin for managing customer reviews and testimonials.

## Docs


- [Documentation](https://myplugin.com/docs/): Full setup and configuration guide

- [FAQ](https://myplugin.com/faq/): Answers to common questions

## Optional


- [Changelog](https://myplugin.com/changelog/): Update history

SaaS Product with API Docs

# MySaaS Platform

> A project management SaaS tool for remote development teams.

## Getting Started


- [Quickstart Guide](https://mysaas.com/docs/quickstart/): Up and running in 5 minutes

- [API Reference](https://mysaas.com/api/): Full REST API documentation

## Blog


- [Engineering Blog](https://mysaas.com/blog/engineering/): Deep dives into how we build MySaaS

## Optional


- [Status Page](https://status.mysaas.com/): Live system uptime and incident history

Agency or Service Business

# Acme Web Agency

> We design and develop high-performance websites and web applications for SMBs in the US and UK.

## Services


- [Web Design](https://acmeagency.com/web-design/): Custom website design services

- [Laravel Development](https://acmeagency.com/laravel/): Backend API and application development

## Portfolio


- [Case Studies](https://acmeagency.com/case-studies/): Real results from real clients

## Optional


- [About Us](https://acmeagency.com/about/): Team background and company history

Bad vs Good llms.txt Example

❌ Bad Example (Too Generic)

# My Website

> Welcome to my website.

## Pages


- [Home](https://example.com)

- [Blog](https://example.com/blog)

✅ Good Example (LLM-Optimized)

# MyPlugin

> A WordPress plugin for managing verified customer reviews with structured schema support.

## Docs


- [Getting Started](https://example.com/docs/start): Setup and installation guide

- [Schema Integration](https://example.com/docs/schema): Improve SEO with structured data

## Blog


- [How Review Schema Works](https://example.com/blog/review-schema): Deep technical explanation

The second example gives clear context, intent, and high-value entry points — which is exactly what LLMs need.

Best Practices

  • Keep it simple and readable
  • Update regularly as AI policies evolve
  • Combine with robots.txt for full control
  • Clearly define licensing terms
  • Avoid contradictions in rules

Limitations of llms.txt

  • Not yet a universally enforced standard
  • Relies on AI companies to respect it
  • No guarantee of compliance

Advanced Tips & Best Practices

  • Keep the file concise (<10KB ideal)
  • Always provide .md versions of key pages
  • Use the “Optional” H2 for skippable content
  • Update regularly
  • Combine with strong internal linking and schema markup
  • Monitor with server logs for AI bot access

Future of llms.txt

As AI regulation grows, llms.txt or similar standards may become widely adopted, possibly enforced through legal frameworks or industry agreements.

Conclusion

Adding an llms.txt file is a low-effort way to speak directly to AI systems. While not a magic bullet, it positions your content for better representation in the age of AI answers. Start simple today and refine over time.

Action Step: Create your file and test by asking an LLM about your site!

Want to Build Smarter Web Applications?

At Codeboxr, we specialize in building scalable SaaS platforms, WordPress plugins, and custom Laravel applications. Let’s turn your idea into reality.

Let’s build the advanced web solution your business demands.