Unity Catalog is an open-source, multimodal catalog platform designed to serve as a universal catalog for data and AI assets. Written primarily in Java, the project is hosted as a sandbox project under the LF AI and Data Foundation and operates under an Apache 2.0 license. The repository provides both an OpenAPI specification and an open-source implementation that enables organizations to catalog, govern, and access diverse data and AI assets through a unified interface.
The platform's core strength lies in its multimodal support across formats, engines, and asset types. It handles multiple data formats including Delta Lake, Apache Iceberg, Apache Hudi via UniForm, Apache Parquet, JSON, and CSV. Beyond tabular data, Unity Catalog extends to unstructured assets and AI models, providing comprehensive governance across the entire data and AI landscape. The system maintains compatibility with Apache Hive's metastore API and Apache Iceberg's REST catalog API, allowing it to integrate with numerous leading compute engines through its open APIs.
The project has cultivated a vibrant ecosystem with support from major technology companies and data platforms. The README lists backing from Amazon Web Services, Google Cloud, Microsoft Azure, Confluent, dbt Labs, DuckDB, Fivetran, and numerous other organizations including NVIDIA, Salesforce, and LangChain. This broad industry support reflects the platform's positioning as a critical infrastructure component for modern data and AI workflows.
The repository's contributor network overlaps with other significant projects including Trino, Microsoft AutoGen, and Apache Iceberg, indicating its integration within the broader data ecosystem.
The platform provides multiple interfaces for interaction including a REST API, command-line interface, and web-based UI. Users can operate on Delta tables through the CLI, integrate with DuckDB for querying, and manage catalogs, schemas, tables, volumes, and functions. The quickstart documentation demonstrates practical workflows such as listing tables, retrieving metadata, and querying data through various tools.
Development infrastructure relies on sbt for building and testing, with requirements for JDK 17. The codebase adheres to Google's Java style guide, verified through automated checks during builds. The project includes comprehensive documentation served through mkdocs and maintains an evolving REST API specification that is currently not assumed to be stable. Deployment is supported through tarball generation for flexible installation options, and Docker Compose support is available for containerized deployments.
