llmfood converts a Docusaurus static HTML build into clean Markdown files optimized for LLM consumption. It:

1. Discovers all pages in a Docusaurus build directory

2. Resolves client-side content that doesn't exist in static HTML (GitHub code references, remote content, mermaid diagrams)

3. Converts each HTML page to Markdown, stripping Docusaurus chrome (breadcrumbs, pagination, TOC, footers)

4. Generates llms.txt

You can check it working with https://docs.sablier.com/concepts/lockup/stream-shapes.md.