Tutorial: Workshop app editor¶
CSA Loom
workshop-appeditor — the Azure-native equivalent of Palantir Foundry Workshop (Loom calls it Atelier): a low-code operational app bound to a Loom Ontology, running on Azure with no Microsoft Fabric workspace required.
What it is¶
A Workshop app is an operational, low-code application whose pages render object views over an Ontology's entity types instead of over a raw database. Widgets read live objects, users traverse links between them, and write-back actions persist changes through the ontology's bound Lakehouse / Warehouse. The app runs on Azure Container Apps over the ontology's existing data bindings — there is no Fabric or OneLake dependency.
When to use it¶
- You have a Loom Ontology (entity types + links + actions) and want a purpose-built app for operators to view and edit those objects.
- You need a screen that lists objects, drills into one, shows related objects across a link, and lets a user submit a change that writes back to the data.
- You want a governed alternative to a hand-built canvas app: every action is recorded as a Thread lineage edge from the app to the ontology.
Step-by-step in Loom¶
- Create the item. In your workspace choose + New item → Workshop app (Fabric IQ category). The editor opens at
/items/workshop-app/<id>. - Bind an ontology. Pick a saved Ontology in the same workspace. Its entity types become the object types available to every widget.
- Add object views. Use Add widget to place an object table (choose the Object type and the properties to show), then chart, metric, and markdown widgets as needed. Paginate rows with Prev / Next.
- Add variables and events. Under Variables, add a variable a widget writes to (e.g. the selected object's key); under Events, wire an event so selecting a row filters a detail widget "filtered by variables".
- Wire a write-back action. Add an action widget, choose the Action kind (create / update), and map its inputs to object properties. The action writes back through the ontology's bound Lakehouse / Warehouse.
- Run an action. Test it in the editor. Loom executes the real backend write and records a Thread edge from the app to the ontology so lineage stays accurate.
The Azure backend it rides on¶
- Compute: Azure Container Apps hosts the running app.
- Data: the ontology's own bindings — an ADLS Gen2 + Delta lakehouse or a Synapse warehouse — so reads and write-back actions hit real Azure data.
- Lineage: Loom's Thread graph records app → ontology edges.
No Fabric required¶
The app binds to a Loom Ontology, not to a Fabric semantic model, and runs entirely on Azure Container Apps over Azure data. No Fabric capacity, workspace, or OneLake shortcut is involved on the default path.
Learn more¶
- Ontology editor tutorial:
editor-ontology.md - Low-code app concepts on Microsoft Learn: https://learn.microsoft.com/power-apps/maker/canvas-apps/getting-started