I read and catalogued 120 published ontologies so that you don't have to. Tell me what you're building, and get back the one — or, more often, the combination — that already covers it, licence terms and all.
There's no vocabulary to learn first. Say what you're modelling the way you'd say it to a colleague — the entities, the relationships, the awkward bits you're not sure how to handle.
Every verified entry opens its real class tree, a worked SPARQL query, and sample data you can click through. You'll know whether it fits in about a minute — not a fortnight.
I read each licence from the primary source and pulled out the parts that actually bite — attribution, share-alike, and whether you're allowed to redistribute it at all.
Plenty of catalogues will happily draw you a plausible-looking class tree. So did this one, for a while — until I went back and checked, and found that most of mine had never been near the actual source file. If a tree wasn't parsed out of the ontology's own published file, it's fiction with good typography.
So now every tree carries its receipts: the source URL, how many classes the file held, how many are shown, when it was fetched, and the exact command that did it. Where a licence forbids redistribution, or the file turned out to be the wrong shape, you get nothing at all — and a note saying which.
50 of 120 entries clear that bar today. The other 70 tell you exactly what's missing, which is the next best thing.
w3.org/ns/prov-o — 32 classes found, 32 shown, retrieved 2026-08-07
Open the tree, filter it, click any class to see just its corner of it.
No hierarchy is served for this entry.
The licence doesn't allow redistribution, so you get a link to the official source rather than my best guess at what's inside.
Metadata, licensing with the operative terms, provenance, and the verified class hierarchies. Static files on a CDN — no key, no rate limit, CORS open. There's an MCP server too, if your client speaks it.
curl -s ontology.guru/api/verified.jsonRead the API docs