Knowledge graph
What Knowledge graph means
A knowledge graph stores things and the relationships between them rather than documents and keywords. Google introduced its own in May 2012, describing "an intelligent model, in geek-speak, a graph, that understands real-world entities and their relationships to one another: things, not strings."
More precisely
The structure is nodes for entities and typed edges for relations, with each node carrying an identifier so it can be referenced from outside. Google's Knowledge Graph Search API returns individual matching entities as JSON-LD, using schema.org types, and states that it "returns only individual matching entities, rather than graphs of interconnected entities." Google now directs new users to Cloud Enterprise Knowledge Graph instead.
Why it matters
Publishers cannot write into a platform's graph directly. What they can do is publish unambiguous, consistent statements about themselves that a graph builder could reconcile against other records.
Example
A knowledge panel is the visible surface of a graph entry, assembled from the entity and its relations rather than from any single page.
Where this came from
Definitions here follow the specification or documentation that defines the term, not a paraphrase of another glossary.
-
Introducing the Knowledge Graph: things, not strings — Google
Google's original description of the Knowledge Graph as a model of real-world entities and their relationships, published 16 May 2012 by Amit Singhal.
Primary source · read 2026-09-03
blog.google/products-and-platforms/products/search/introducing-knowledge-graph-things-not/
-
Google Knowledge Graph Search API — Google
The API returns only individual matching entities rather than graphs of interconnected entities, uses schema.org types and JSON-LD, and new users are directed to Cloud Enterprise Knowledge Graph.
Primary source · read 2026-09-03