Most developers talk about unifying data. But in reality, data lives everywhere — in lakes, warehouses, databases, streaming systems, AI/ML pipelines. Trying to centralize or replace them is costly, slow, and often fails.

So instead of asking “How do we unify the data?” we asked: “Can we unify the metadata?”

That’s the idea behind Apache Gravitino — an open-source “catalog of catalogs” that sits above your existing systems and provides:

Unified metadata governance without replacing your stack

Federated access to diverse systems (SQL, NoSQL, lakehouses, ML/AI)

A lightweight, extensible platform you can contribute to and extend

Website: Datastrato

Code: Apache Gravitino on GitHub

We’d love feedback from HN: Does focusing on metadata instead of data solve a pain you’ve seen? What gaps do you think still exist in the “data & AI catalog” space?