Status

What is built today, and what is not yet.

This page is kept factual and dated. Last reviewed 27 Oct 2026.

Built and running

• One domain, static hosting on Cloudflare Pages, built by GitHub Actions from one repository.
• 41 sections in the registry, each on its own branch or snapshot branch, each under its own path, all reachable from the header selector.
• One core catalogue database rebuilt on every deploy, read in the browser with SQL and GraphQL, with content-addressed storage so a shared file is stored once.
• Snapshot branches pinned by sections.lock for 7 sections with live origins, so builds no longer depend on those origins answering.
• Whole-database export at /db/export/, including one slice per section.
• Agent material kept out of human view (metadata, /agents.json, /llms.txt, /academy/, and the build pass for self-describing notes).

Not yet

• Words in templates. Most sections are full static sites whose text sits in their HTML. The core database indexes those pages; it does not yet generate them. Sections generated from data today are listed in Front ends.
• Own database per section. The convention exists (public/_db/section.sqlite, merged into section_records) and one section uses it: nft-art. The other sections keep their data as files on their branch, and a few with live features keep a Worker store; their export slices are derived from the core database.
• Crawled sections. 4 sections (email, dartboard, wales, terrain) are still copied from their live origin on each build.
• Data still read from the old repository. The dartboard's roadmap data is published from a branch of the earlier repository before it reaches this site.
• Branch protection. The section and main scope checks run in CI; repository rulesets that enforce them for every push are not switched on yet.

Order of work

1. Snapshot the remaining crawled sections into section/<id> branches.
2. Move the dartboard's data source into its own section branch.
3. Move section text into data files, starting with the sections that are already mostly data (e-books, nomenclature, records registers, title chain), then the evidence and research pages.
4. Give each section its own section.sqlite in the nft-art convention, so its pages read their words from section_records.
5. Turn on repository rulesets for main and the section branches.