Introduction 

What is llms.txt? 

llms.txt is an emerging standard for providing AI and Large Language Model (LLM) crawlers with structured information about a website. It serves as a machine-readable index that helps AI systems understand your website's content and structure.

The llmstxt.org specification defines how websites can provide this information in a standardized format.

Concept 

This extension provides a two-tier approach for LLM content access:

  1. llms.txt Index File - One file per language containing:

    • Website metadata (title, description, domain)
    • Page structure with SEO descriptions and keywords
    • Instructions for accessing full page content

  2. Markdown Content Access - Access any page content via:

    • .md suffix - Returns clean Markdown with YAML frontmatter
    • Example: /about.md returns the "About" page as Markdown

Multi-Language Support 

  • One llms.txt per language - Every enabled language has its own llms.txt below its base, e.g. https://example.com/en/llms.txt, with the pages in this language and links to the llms.txt of the other languages.

  • Language-specific content - Access any page in any language using the .md suffix with language URL prefix:

    • Default: https://example.com/about.md
    • English: https://example.com/en/about.md

Features 

Core Features 

  • Automatic llms.txt generation with smart caching
  • Markdown output for all content pages via .md URL suffix (can be disabled per site)
  • Multi-language support: one llms.txt per language, linked to each other
  • API key protection for restricted access
  • YAML frontmatter in Markdown output with page metadata

Page-Level Control 

  • LLM tab in page properties for fine-grained control
  • Custom descriptions and summaries per page
  • Keywords for better LLM understanding
  • Priority setting (0-100) for page ordering
  • Exclude option to hide specific pages from llms.txt

Technical Features 

  • 24-hour caching for optimal performance
  • HTML-to-Markdown conversion using League/html-to-markdown
  • Clean output - removes scripts, styles, navigation elements
  • UTF-8 BOM in the Markdown output for proper encoding detection
  • Backend notification if robots.txt lacks llms.txt reference
  • Header link injection (<link rel="alternate">) in HTML pages

Requirements 

  • TYPO3 13.0 - 14.x
  • PHP 8.2 or higher

Supported Content Elements 

The extension converts the following TYPO3 content elements to Markdown:

  • Header (header)
  • Text (text)
  • Text with Image (textpic, textmedia)
  • Image (image)
  • Bullet List (bullets)
  • Table (table)
  • HTML (html)
  • Menu elements (menu_*)
  • All other elements via HTML-to-Markdown fallback