How bulk access works
Bulk access is a request, not a download. Each request is reviewed by the project before any extract is shared, so that we know how the data are being used and can supply the right version of it.
- Tell us who you are and where you are based.
- Describe what you intend to do with the data.
- Say which part of the registry you need, and for what period.
Browsing and searching the registry needs no request and no account.
Coverage and validation are still in progress, so it should not yet be interpreted as systematic or fully validated.
License and reuse
Whether you take the data through this request, through the API, or from the site itself, the same terms apply.
- Data — CC BY 4.0
- Share it and build on it, including commercially, as long as you credit the project and say what you changed.
- Code — MIT
- The site and the tooling behind it are open source, and reusable with the copyright notice kept intact.
Suggested credit: Global Government GenAI Observatory (G3O), 2026. g3observatory.org, CC BY 4.0. Please cite the dataset version and the date you accessed it.
A published version is fixed and is never overwritten. If a paper cites a version, ask for it and the project will supply that version, so a published result can be checked against the data it was actually run on. Archival deposits with DOIs will follow in Stanford Dataverse after each completed collection sweep.