Explore, search and annotate the datasets in a rask lakehouse, and run model pipelines over them. The workbench is a client of rask: it reaches a lakehouse only through rask's published API and the credentials its catalog vends.
- explorer: the viewer, search and the explorer UI
- annotator: the annotator, its assist runner and its UI
- flows: model pipelines and the studio UI
Extracted with its history from rask at 64ef118a9 (2026-11-14); see issue #1 for the plan. Development: make install,
then make check and make test. Images are built with Dagger: dagger call image --name=<stem>.