The register is published to be quoted, cited, and built on. The whole thing is one JSON file. No key, no sign-up, no rate limit.
One file, all 29 entries, rebuilt whenever an entry is verified.
Each record carries the fields below. The file itself is a schema.org
Dataset node with a dateModified stamp, so it
can be cited as a dataset, not just linked to.
Each register entry page carries the same data inline as schema.org
JSON-LD in the <head>, as two nodes: the instrument
itself (Legislation for statutes, regulations, and
executive orders; CreativeWork for standards and
advisories), identified by its own primary-source URL, and AI Reg.
Index's page about it, a separate CreativeWork published
by AI Reg. Index. The two are kept distinct so nothing implies AI Reg.
Index published the underlying law. The register itself carries a
Dataset and an Organization node; entry pages
add a BreadcrumbList. An answer engine or a scraper can read
the register without touching the JSON file.
With 29 records, filtering by jurisdiction, standing, or tag is a one-line operation on whatever consumes the file.
Register text and data are licensed CC BY 4.0 — quote it, reformat it, build on it, with attribution: AI Reg. Index, airegindex.com, plus the date you retrieved it. The primary documents themselves belong to the governments and standards bodies that publish them.