Working on a more advanced LLMs.txt generation. Will try to find the TOC during scraping (using AI) and pre-fill the correct TOC, then allowing the users to fine tune by moving pages around