For agents

This site is written for people and for software agents from one source. Every page has a human view, a machine view and a Markdown twin, and all three carry the same words.

Illustration: A map sheet lifting away from the page to show the field book underneath. The same lines run through both, drawn on the map and written in the book. A metaphor for the human view and the machine view of one page.

Machine files

  • llms.txt: an index of every page, one line on each.
  • llms-full.txt: the Markdown of every page in one file.
  • facts.json: every fact the site states, with its record id and source.
  • sitemap.xml: the canonical address of every page.
  • robots.txt: the crawl policy and the content signal.
  • A Markdown twin of every page. Add .md to the address. The home page twin is index.md.

How to read it

  • Treat the facts file as the list of what the site claims.
  • Read launch_state in the front matter of any twin. While it is pre_launch, every capability described is a design, not a running service.
  • Cite the canonical page address, not the twin. Each twin names it in its url field.
  • Pages marked status planned describe something that does not exist yet.

What is not here

  • No API. The docs keep a reserved place for product reference.
  • No forms and no write paths. The site only serves files.
  • No cookies, no analytics, and no requests to other sites.

The detail

The machine view, explained in the docs.

Last reviewed .