llms.txt and RSS: Discovery Outputs Built From Your Content
Featurellmsrsssitemapseo

llms.txt and RSS: Discovery Outputs Built From Your Content

Site-wide and per-collection llms.txt, RSS and sitemap files are generated from the same content, including live D1 rows.

By

llms.txt and RSS outputs

Every collection can publish three machine-readable outputs without extra pages to maintain.

What is generated

  • /llms.txt lists the site-wide content that agents and answer engines can read.
  • /<collection>/llms.txt does the same for one collection, for example /docs/llms.txt.
  • /rss.xml is the combined feed and /<collection>/rss.xml a feed per collection.
  • /sitemap.xml and the taxonomy pages cover discovery by crawlers.

Controlled per entry

Front matter decides what is exported:

index: true
sitemap: true
rss: true
llms: "full"   # or "basic"

Drafts and entries with these flags off are left out. Collections that read from Cloudflare D1 are included once their rows are published, because the exports use the same source-aware resolver as the pages.

Written for readers

Good entries have stable URLs, a clear title, a plain summary, useful headings and taxonomy fields. The exports reuse that structure, so improving the page improves the output.

Try them on this site: /llms.txt, /rss.xml, /sitemap.xml.