📜 docusaurus-plugin-llms

A Docusaurus plugin that writes llms.txt, llms-full.txt, and Markdown copies of your docs, following the llmstxt standard.

During a production build, this plugin scans your Docusaurus docs and writes an llms.txt file linking every documentation page plus an llms-full.txt file bundling all your content into one document, so LLMs can read your docs without parsing HTML. It needs no configuration.

Features

• Generates llms.txt with section links and llms-full.txt with all content in one file.
• Optionally writes individual .md files per page, which the llmstxt.org spec recommends.
• Controls document order and transforms URLs with glob patterns.
• Cleans content for LLMs: strips HTML, removes import statements, and drops duplicate headings.
• Rewrites relative image paths to absolute build-output URLs.
• Supports custom LLM files, multi-instance docs, and multi-version output.

Getting started

Install the plugin as a dev dependency:

npm install docusaurus-plugin-llms --save-dev

Register it in docusaurus.config.js:

module.exports = {
  // ... your existing Docusaurus config
  plugins: [
    'docusaurus-plugin-llms',
    // ... your other plugins
  ],
};

On your next npm run build, the plugin writes llms.txt and llms-full.txt into the build output. To check the output, see installation. To change the defaults, see configuration options.

Documentation

The docs are published at rachfop.github.io/docusaurus-plugin-llms, a Docusaurus site that runs this plugin on its own documentation. Its llms.txt is the plugin's live output.

• Overview: what the plugin generates, when it runs, and where page titles and descriptions come from.
• Installation: install the package, register it, and check the output.
• Configuration options: the full reference for every option the plugin accepts.
• Generating content: multiple doc sections, custom root content, custom LLM files, and partials.
• Generating individual Markdown files: what generateMarkdownFiles produces and how files are named and laid out.
• Content cleaning: strip HTML, remove imports, drop duplicate headings, and rewrite image URLs.
• Ordering and path transformation: control document order, transform URLs, and tune batch processing.
• Multi-version output: publish a separate set of LLM files per documentation version.
• Best practices: option settings for API reference, tutorials, and multi-language docs.

License

This project is licensed under the MIT License.