dbt Integration

dbt Integration

Connect your dbt project once and every model becomes a governed cube. The dbt integration converts dbt models into cubes — dimensions, measures, column descriptions, primary keys, and joins inferred from dbt relationship tests and foreign-key constraints — so you never redefine your transformations by hand, and dbt stays the single source of truth.

Sync however fits your workflow: pull manually from the IDE, call the REST endpoint from CI right after dbt run, or trigger on every push with a webhook. Every automated sync lands on a review branch that a human approves before it reaches production. Cube connects to your repository over HTTPS or SSH deploy keys and only ever parses your project — it never touches your data warehouse. Read more in the announcement.

Other improvements and fixes:

Workbooks

• You can now drill down on a time dimension in Workbook Table charts and the results grid
• Introduced a new-tab experience that lets you quickly start an exploration from a single launchpad: browse or search your semantic model and source tables, or duplicate any existing workbook tab

Dashboards

• Read-only viewers of a published Dashboard can now subscribe to and unsubscribe from its scheduled refresh notifications
• Scheduled refresh email notifications can now be sent to user groups

Visualization

• Added a unified, context-aware color selector shared across chart series, maps, and conditional formatting
• Fixed chart dates not matching the results table by anchoring chart time values to UTC

Administration

• Added a tenant-wide setting to restrict CSV downloads, plus a deployment-scoped Download Data permission for custom roles
• Added a deployment-scoped Manage secrets permission to grant access to one deployment's configuration and credentials without exposing others
• You can now select historical Cube versions in any release channel

AI

The Cube MCP server can now edit the semantic data model, not just read it — external agents can open a dev branch, write model files, and commit changes for review.

Embed

Creator Mode embeds can now hide the Semantic SQL and Generated SQL views, so embedded end users can't see your data source's table and schema names.

Integrations

• You can now search reports and folders by name in the Google Sheets add-on
• Users with access to a shared report can now refresh and place it in the Google Sheets and Excel add-ons, even if they didn't create it