OpenCyc
OpenCyc is the freely released portion of Cyc, a project started in 1984 to write down the everyday common sense that people never bother to state out loud: that dropped things fall, that a person cannot be in two places at once, that Paris is a city and a city is a kind of place. It ships as a formal ontology of hundreds of thousands of concepts, which Cyc calls collections and individuals, wired together by millions of assertions built from predicates and a deep hierarchy of types.
The last public release landed in 2012 as downloadable OWL files. You load those into a triple store or an RDF library and query them much as you would any other knowledge graph. There is no live API and no ongoing maintenance, so treat it as a fixed snapshot rather than a service. Because cyc.com is now Cycorp's commercial site, the original OpenCyc downloads can take some tracking down, but archived copies remain available.
For RAG, its value is as a scaffold of concepts and types rather than a well of facts to retrieve. It earns its keep if you want to tag entities against a principled category hierarchy, add common-sense typing to your pipeline, or study how hand-built symbolic knowledge is organised.
The caveats are real. It is frozen, so nothing reflects the world after 2012, and its instance data, meaning actual named things, is thin next to a live source like Wikidata. It is also only a slice of full Cyc: the inference engine and the richest rules stay proprietary. The redeeming feature is the licence. Apache 2.0 is genuinely permissive, so commercial use is fine with no ShareAlike or attribution strings attached.
Reach for Wikidata or DBpedia when you need broad, current, well-populated facts, and ConceptNet for lighter common-sense links. OpenCyc complements those with formal structure rather than scale.
Related sources
DBpedia
A knowledge graph built by pulling the structured parts of Wikipedia, mainly the infoboxes, into machine-readable data. It holds billions of facts about people, places, organisations, and more as RDF triples, small subject, predicate, object statements, which you can search with the SPARQL query language.
Freebase
A collaborative knowledge base once run by Google, now retired but still available as downloadable data dumps. It holds structured facts about millions of entities, and much of its content has since moved into Wikidata.
Structured Wikipedia
Wikipedia rendered as pre-parsed, machine-readable JSON: abstracts, short descriptions, infoboxes, sections, parsed tables and references, with links to Wikidata entities. A beta from Wikimedia Enterprise covering nine languages, also mirrored on Hugging Face. The section-segmented shape a RAG pipeline actually wants.
SYNTH
A fully open synthetic corpus of amplified multilingual encyclopaedic text with built-in reasoning traces and exercises covering RAG, information extraction and QA. Released by PleIAs with the AI Alliance, it targets training and evaluating small, grounded, citeable reasoning models.