llms.txt works best as an index of important links, not as a second copy of the catalog.
Your sitemap already lists URLs for crawlers. llms.txt lists the policies, pages, collections and every active product you include on that task.
What stays out
- The model is told not to invent extra pages
- There is no product cap
- You can leave products unchecked if you only want policies and pages
If you need every URL in the store indexed for search engines, that job stays with sitemap.xml. LLMs TXT checks that file. It does not try to replace it.