Generate LLM-friendly documentation indexes offline instantly. Structure tags, categories, and references for AI crawlers.
⚙️ Complete the configurations below to generate your llms.txt structure.
Project Metadata
Documentation Sections
Generated File Output
Full Mode
What is the LLMs.txt standard?
The **llms.txt standard** is a recently proposed web convention (hosted at the root `/llms.txt` of a project domain) that serves as an index map for large language model crawlers, agent systems, and code assistants. Much like how a traditional `sitemap.xml` guides Googlebot through web hierarchies, `llms.txt` provides clean, markdown-based paths that help AI systems crawl and read your site's codebase, guidelines, APIs, and articles.
A primary `llms.txt` outlines the core project features and links to primary pages, while a secondary `llms-full.txt` compiles detailed documentation files and full page codeblocks for models to ingest in single-prompt context windows.
Frequently Asked Questions (FAQ)
Where should the llms.txt files be hosted?
The compiled `llms.txt` file must be uploaded to the **root directory** of your website server, making it accessible at `https://yourdomain.com/llms.txt`. If you generate a secondary full index, it should be saved as `llms-full.txt` and referenced as a link inside the main `llms.txt` file.
What is the difference between llms.txt and sitemap.xml?
A `sitemap.xml` is structured in rigid XML schema tags specifically for traditional indexing spiders to discover URL paths. An `llms.txt` is structured in human-readable Markdown syntax. It includes taglines, summaries, category headers, and descriptions for each link, helping LLMs understand the semantic context and relevance of pages before reading them.
Why do AI code assistants prefer this file?
When systems like Claude, ChatGPT, or Antigravity fetch documentations, they read web pages. If a website contains bloated HTML layouts, navigation menus, and script tags, it consumes valuable LLM token budgets. The `llms.txt` convention lists direct routes to lightweight Markdown pages, enabling models to fetch exact docs instantly with zero parsing overhead.