Amazon SageMaker Unified Studio gets custom asset types for cataloging anything!
Hi, it's Shiichan! Today I've got a handy data-catalog update from AWS's SageMaker Unified Studio.
AWS What's NewWhat was announced?
According to AWS's What's New, Amazon SageMaker Unified Studio now supports custom asset types in IAM-based domains. With this, domain administrators can catalog any format of asset, things like medical imaging files sitting in Amazon S3, revenue dashboards built in PowerBI, or PDF research reports generated by a third-party platform, all inside the SageMaker catalog.
The story so far
Before this, assets that didn't fit the catalog's existing formats had to be tracked with separate tools or processes, team by team. With custom asset types, everything gets pulled into the same catalog regardless of its underlying format, so teams can search, discover, and subscribe to it all in one place.
What changes
Administrators start by creating a custom asset type with a name, description, and optional metadata forms defining the fields each asset should carry. Individual assets are then created from that type, enriched with glossary terms and README documentation to add business context for both humans and AI agents. Once published, anyone in the domain can find the asset by name, type, or glossary term and request a subscription through the same governed workflow used for every other catalog asset.
Dive Deep
This is available in every AWS Region where Amazon SageMaker Unified Studio runs. If you have admin access and want to try it, the SageMaker Unified Studio user guide walks through the setup.
Wrap-up
- SageMaker Unified Studio's IAM-based domains now support custom asset types
- Mismatched formats like medical images, BI dashboards, and PDF reports can now live in a single catalog
- The flow: admins define a type, create and publish assets from it, then anyone in the domain can search and subscribe
- Available in every supported AWS Region
- Great for data platform teams and admins who want centralized management across varied data assets