Anzubo

Open data

Everything this site says about a robot comes from a record, and the records are published as files you can download and reuse: every robot in the catalogue, and every robot examined and not listed, with the source behind each fact.

The files

File What it holds Records
/data/robots.json Every robot that meets the criteria, complete: what it is, what it is claimed to do, where it can be had, at what published price, and every source. 8
/data/robots.csv The same robots, one row each, with the fields most comparisons need. For a spreadsheet. 8
/data/not-qualified.json Every robot examined and not listed, complete: which criteria it fails and passes, the reasons, what would change the outcome, and every source. 11
/data/not-qualified.csv The same robots, one row each. 11

These addresses are stable: they will keep working, and always hold the current data.

Licence

The data is published under Creative Commons Attribution 4.0 (CC BY 4.0). You may copy, share and adapt it for any purpose, including commercially, as long as you credit it as Anzubo, anzubo.com and link to the licence.

The licence covers this site’s own work. It does not cover images: each photograph has its own licence and author, given beside it in the data and on the page, and is reused on its own terms. Nor does it cover the vendors’ own words, quoted verbatim in capabilities[].claim and global_release_statement.text: those remain theirs.

Tracing a fact to its source

Every fact in the JSON files carries evidence: a list of ids, each naming an entry in the same record’s sources, with its address, publisher, kind, language and the date it was retrieved. That is the same link a reader of a robot page follows from a fact to its source. An empty list means there was nothing to cite (typically a criterion failed because nothing has been published), and the note beside it says what was looked for.

Reading it correctly

Which version this is

Every file says when it was generated and from which version of the dataset (generated and source_commit), and carries the licence and attribution itself (in the JSON at the top, in the CSV at the end of every row), so a copy can always be traced back.

The files on this site were generated 2026-10-04T09:58:22Z, from version fa3afa74d867a4a60820f8cfa2ad2d0c54b2e96c.

About the CSV files

UTF-8, comma-separated, one row per robot, with a header row. Where a cell holds several values they are separated by “; ”. The JSON files are complete; the CSV files hold the common fields only.

Nothing here has been tested hands-on. The records are researched and kept up to date by AI agents from published sources, as Method explains.