First off, congratulations on the milestone!

This seems like the first viable opensource alternative to commercial Data Catalogs from Alation, Collibra, Waterline, Informatica, etc.

The key differentiator here seems to be the modular and scalable architecture, combined with the consistency of the core graph model and exetensibility of graph models in general, which should allow contributions of different parts of the metadata through over time.

In particular, it seems like this could allow the creation of something similar to Alation’s Behavior I/O off of the data generated from CDC tools in the Confluent community, which could be an industry disruptor. In order to get to that point, you would also need to have a good lineage ingestion & inference solution that doesn’t fully rely on Airflow.