Does this plugin generate llms.txt from HTML

[noklam]

I read the source code and it seems like the plugin processes HTML and parse the content. Why does it needs to parse HTML instead of using the raw .md file that was used to process the HTML?

[pawamoy]

Hi @noklam , that's a good question 🙂

I tried to explain that in the README. With MkDocs, there are plenty of plugins and extensions that allow injecting contents in a document. These injections only happen when converting the Markdown content to HTML. If we want these injected contents to appear in the final .md pages, we therefore have to rebuild the llms.txt pages from HTML.

I hope that answers your question!

[noklam]

Thanks! This make sense. I think for my use I only need the basic markdown but the codebase is very useful as a reference implementation.

[pawamoy]

We could consider having an option to decide whether to use raw markdown directly or convert back from HTML. Feel free to open a feature request 🙂

[noklam]

Right now I am trying to implement this as a separate plugin, it's also for my personal learning to get familiar with mkdocs plugin system. If it end up making more sense to live in the same repo I will raise a PR.