DBpedia Launches a Knowledge Graph Catalog

The DBpedia Association published the inaugural release of its Knowledge Graph Catalog on November 9. It's a standardized metadata registry covering 24 knowledge graphs sourced from DBpedia Databus and DBpedia MOSS, shipped as a downloadable snapshot (metadata about the graphs, not the underlying datasets themselves) and tagged on GitHub as kg-catalog-2026-11. DBpedia plans to release updates every month as new knowledge graphs and metadata come in.

Three metadata layers, not one flat list

The release is organized into three collections rather than a single spreadsheet:

• Source metadata: the original per-graph YAML descriptions contributors submit.
• Release metadata: RDF describing 745 individual knowledge graph versions over time.
• Graph-level metadata: profiles rolling those versions up for the 24 catalogued graphs.

This three-tier structure lets the catalog track a knowledge graph's evolution: which version introduced which vocabulary, when a dataset was last refreshed. Most registries of this kind only list a current snapshot and skip that history.

Why a catalog is a data integration problem, not just a directory

A metadata catalog is only useful if an organization can actually query it alongside its own data. That means joining "which knowledge graphs expose a FOAF profile" against an internal data inventory, or scripting a monthly diff against the DBpedia release feed to flag when a dependency changes. Both are integration jobs: mapping an external, RDF-shaped metadata format onto internal systems that mostly weren't built with RDF in mind.

Making the catalog a living input to an internal data platform, instead of a page someone checks manually once a quarter, takes an integration layer that's actually built and maintained. A disciplined custom software development process is what gets teams there — treating external metadata feeds as a dependency to design around rather than an afterthought.

What to watch next

DBpedia's stated cadence is monthly, so the first real signal will be whether the December release lands on schedule and how many new knowledge graphs get added beyond the initial 24. Teams building on CCO or BFO-conformant ontologies may also want to check whether any of the 24 cataloged graphs end up exposing CCO-aligned or ISO/IEC 21838-conformant metadata as the catalog grows. The source metadata collection is open to contributors, so that's worth checking against each monthly update rather than just the November snapshot.