This VitePress plugin automatically generates LLM-friendly documentation, transforming standard website content into raw Markdown files optimized for efficient processing by large language models. It targets developers using VitePress who need to make their documentation accessible and parseable by AI, offering a streamlined way to create machine-readable knowledge bases. The primary benefit is enabling LLMs to understand and utilize documentation more effectively without the overhead of HTML or complex formatting.

How It Works

The plugin integrates seamlessly into the VitePress build process. It scans documentation files and generates two key outputs: llms.txt, which provides a structured list of links to different documentation sections, and llms-full.txt, compiling all content into a single, contiguous Markdown file. It supports custom content filtering using <llm-only> tags for LLM-specific instructions and <llm-exclude> tags for human-only content, adhering to the llmstxt.org standard for LLM documentation best practices.

Quick Start & Requirements

• Installation: npm install vitepress-plugin-llms --save-dev
• Prerequisites: VitePress, Node.js environment.
• Usage: Add llmstxt() to the vite.plugins array in your .vitepress/config.ts configuration file.
• Links: npm package, GitHub repository

Highlighted Details

• Zero-configuration setup with VitePress, works out-of-the-box.
• Supports content generation for LLMs, including optional "Copy/Download as Markdown" buttons for users.
• Allows fine-grained control over LLM-specific content using <llm-only> and <llm-exclude> tags.
• Customizable through plugin settings, including file ignores, custom templates, and variable injection.
• Frontmatter descriptions enhance the clarity of links in the generated llms.txt file.

Maintenance & Community

The project is actively maintained, with clear channels for bug reporting and feature requests via GitHub Issues. Sponsorship is encouraged to support ongoing development. While specific community channels like Discord or Slack are not listed, the project follows standard open-source contribution guidelines.

Licensing & Compatibility

The plugin is released under the MIT License, which permits broad use, including commercial applications and integration into closed-source projects without significant restrictions.

Limitations & Caveats

The plugin explicitly states that it is intended for English-language documentation only, advising against its use for repositories with documentation in other languages. Experimental features are available, and their stability may vary.