Skip to Content
Open-source data governance

Governance beyond Databricks: an alternative to Unity Catalog

Unity Catalog governs the Databricks lakehouse well. Classifyre governs everything around it too: on-premises databases, file shares, SaaS tools and lakehouse files, with the same masks whether people query in SQL, Excel or Power BI.

Query data
crm · PostgreSQL
lake · Parquet in S3
▶ Run
SELECT c.email, c.city, sum(o.amount) AS revenue
FROM crm.customers c
JOIN lake.orders o ON o.customer_id = c.id
GROUP BY 1, 2;
3 rows · 2 sources joined · email hashed
emailcityrevenue
5fc7613cacc41469Graz4,812.40
8169dd18e1635846Linz1,130.00
8a84971c7a971d8aWien2,977.15
A Databricks Delta table joined with an on-premises database, masks applied.
How it works

Where Classifyre is different

  1. 01

    Across systems, not one platform

    40+ source types under one set of grants, masks and audit, including the Databricks Delta tables themselves.

  2. 02

    Masks without writing functions

    Masks follow the classification: redacted, last four, hashed, empty or hidden, chosen in a dialog rather than written as SQL functions per column.

  3. 03

    Discovery included

    The personal data is found by the same product that masks it, in tables and in documents.

In plain words

When each one fits

If your data and your users live in Databricks, Unity Catalog is the natural governance layer, and its row filters and column masks work inside Databricks compute.

If your estate is wider than one platform, or must run on your own servers, Classifyre gives one governance layer across it and reads Databricks Delta tables as one source among many.

Side by side

Classifyre and Unity Catalog side by side

As of November 2026, from each project's public documentation.

AspectUnity Catalog (Databricks)Classifyre
LicencePart of the Databricks platform; an open-source Unity Catalog server (LF AI & Data) offers a subsetApache-2.0, every feature
ScopeThe Databricks lakehouse and engines that use its APIs40+ source types: databases, lakehouses, files, SaaS tools
Column masksSQL functions per column, or tag-based ABAC policiesChosen per column or by classification: redact, last four, hash, empty, hide
Row filtersSQL functions per table, or ABAC policiesPer group, failing closed
Finding personal dataData classification for Databricks tablesDetectors across tables and documents, with scores and reasons
DeploymentDatabricks cloud serviceSelf-hosted: Docker, Kubernetes, OpenShift; runs offline

Databricks ships changes often; check their documentation for the current state. Corrections welcome at [email protected].

What it does not do

Classifyre does not enforce its masks inside Databricks compute. Notebooks and jobs that read Delta tables directly in Databricks follow Unity Catalog's rules, not Classifyre's.

Your move

Run it on your machine tonight

One Docker image with the database, the UI and the workers in it. No signup; sign in as classifyre / classifyre. Your data stays with you.

docker run -d --name classifyre \  -p 3000:3000 \  --shm-size=1g \  -v classifyre-pgdata:/var/lib/postgresql/data \  -v classifyre-data:/var/lib/classifyre \  -v classifyre-uv-cache:/cache/uv \  classifyre/all-in-one:0.9.0
Setup and configuration →
helm install classifyre \
  oci://registry-1.docker.io/classifyre/classifyre-core \
  --version 0.9.0
Helm →
Or start with a pilot

Four to six weeks on your data: 3 to 5 sources, detection tuned, masking on, one BI connection, and a findings and access report. Fixed scope, fixed price.

Talk to us about a pilot
Straight answers

Questions

More detail lives in the docs · docs.classifyre.com