YAGO is a knowledge base built by combining Wikipedia's facts with WordNet's taxonomy of concepts and GeoNames' geographic data. The payoff is a graph that pairs broad coverage with an unusually clean type system: every entity is slotted into a consistent hierarchy of classes, and the facts are checked for logical consistency, so the error rate stays low next to graphs scraped straight from raw text. It holds millions of entities and a much larger body of facts, each written as an RDF triple, a short subject, predicate, object statement such as "Marie Curie, wonPrize, Nobel Prize". A distinctive touch is that many facts carry time and place annotations, recording when and where they held true.
There are two practical ways to work with it. You can download the dumps in Turtle or N-Triples and load them into a triple store such as Apache Jena, GraphDB, or Blazegraph, or you can query the hosted SPARQL endpoint, SPARQL being the query language designed for knowledge graphs. For retrieval you rarely hand raw triples to a language model. The usual pattern is to flatten the facts about each entity into short natural-language sentences and embed those, or to keep YAGO as a structured store your pipeline queries to check and enrich generated answers.
YAGO earns its place when precision and clean entity types matter more than raw coverage: entity linking, telling apart people who share a name, grounding claims against dependable facts, or feeding a graph-aware (GraphRAG) pipeline. The time and place annotations make it a good fit for questions about who held which role, and where, over time.
The trade-offs are straightforward. YAGO is smaller than Wikidata and refreshed far less often, so it trails on recent events, and you need to be comfortable with RDF and SPARQL rather than plain files. Licensing is easy: it ships under CC BY 3.0, so commercial use is fine as long as you credit YAGO. Reach for Wikidata when you need live breadth, DBpedia for another Wikipedia-derived graph, and YAGO when you want the cleanest, best-typed facts.
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.
OpenCyc
The open release of Cyc, one of the oldest attempts to hand-build common-sense knowledge for machines. It holds hundreds of thousands of concepts and millions of assertions about how the everyday world fits together. Now archived, but the data is still available.
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.