Data and reuse

The catalog behind this site is published as static files you can download and reuse. There are no accounts, keys, or sign-ups. The files are rebuilt whenever the site is published, so check the build time inside each file.

Files

• /data/catalog.json: every published organization, open artifact, news item, and People Behind Local SI profile, with sources and dates. Also schema_version, rubric_version, build_at, and catalog counts.
• /data/search-index.json: the compact index behind site search, with the name, type, summary, keywords, and page link for each published record.
• /data/jobs.json: currently open job postings from the employer feeds the site covers, with when each was last seen.
• The State of American AI: each edition's figures as CSV and JSON, for example /report/2026-10/data.csv and /report/2026-10/data.json. Editions are frozen when published, include the definition of every figure, and are CC BY 4.0.
• /news/feed.xml: Latest news as an RSS 2.0 feed.
• /sitemap.xml: every public page.
• USASI catalog on Hugging Face (external site: huggingface.co): a dated snapshot of the published organizations, open models and tools, and news as CSV tables, with a dataset card. It can be loaded with the datasets library. The files on this site are always newer; People Behind Local SI profiles and job postings are not included there.

Downloading a selection

In Companies & Labs and Open Models & Tools, filter the list and use Download these to save exactly the records shown, as CSV or JSON. Each row keeps the record's page link and its last review date, and the file records the filtered view it came from and the snapshot date, so others can trace every row back to its sources. The file is built in your browser; nothing is sent to USASI. The comparison workspace offers a similar export.

What is included, and what isn't

• Only published records are exported. Drafts and records under review never appear, and links from published records to unpublished ones are removed, so an export never reveals that an unpublished record exists.
• Each artifact includes a computed_tier calculated with the published openness rubric. It describes documented availability under this catalog's rules; it is not a certification or an external definition.
• Counts describe this catalog's coverage. They are not a census, a ranking, or a measure of capability.

Identifiers

Each record has a slug, which is also its page address (for example /companies/<slug>/ or /open/<slug>/). Within a record, every source has an id, and each claim lists the source IDs that support it in source_ids. Use the slug as the stable identifier when linking to or joining against a record.

What the dates mean

• updated_at: the last substantive edit to the record.
• last_reviewed: when the record's evidence as a whole was last checked.
• reviewed_at on a claim or source: when that specific item was checked.
• accessed_at on a source: when it was read; an automated fetch is not an editorial review.
• released_at: the documented release date of an artifact, at the precision the evidence gives.
• build_at: when the file was generated. A rebuild never changes any of the editorial dates above.

The methodology explains these in more detail.

Reuse terms

Original catalog prose (summaries, explanations, notes, and editorial pages) is licensed under Creative Commons Attribution 4.0 International (CC BY 4.0) (external site: creativecommons.org). Suggested attribution: "Source: USASI catalog (unitedstatesofamericasuperintelligence.com), CC BY 4.0." Facts such as names, dates, license identifiers, and URLs are not claimed as copyrighted.

That license covers this project's own writing only. Models, software, datasets, and documents described in the catalog keep their own licenses, which each record links to. Names, logos, and marks belong to their owners. Job postings are factual fields and links taken from employers' own feeds; they are not relicensed by this project.

Corrections

If you find an error in the data, report a correction with the record, the field, and a supporting primary source. The changelog records catalog and policy changes.