Data Apps

Data Apps: Storage Access via Query Service

We've shipped Storage Access for Data Apps — a first-class way to read and write Keboola Storage tables from within a running app, without routing data through input mapping or redeploying on every change.

Architecture

When enabled, Keboola provisions an ephemeral workspace per app lifecycle (deploy, redeploy, wake from sleep). That workspace receives scoped permissions (SELECT, INSERT, UPDATE, DELETE, TRUNCATE) on the tables you configure. All SQL goes through the Query Service, which handles auth, billing tracking, and automatic metadata refresh after writes.

What you can build

  • Live dashboards that always reflect the latest data (no snapshot lag)
  • Data entry / correction UIs that write directly back to Storage
  • Approval and review workflows over existing Storage tables
  • Multi-user collaborative tools (with app-level concurrency handling)

Important notes

  • Snowflake only; BigQuery support planned
  • Query Service accepts raw SQL — validate and sanitize all user input before interpolation. First-class parameterization helpers (SQL.literal(), SQL.ident()) are in development
  • Workspace credentials rotate on each app restart; do not persist WORKSPACE_ID across sessions
  • Permission changes require app redeploy to take effect

Setup guide and code examples →