Databricks Apps
You'll learn how Databricks Apps let you build custom web applications that run on Databricks with native auth and data access in ~5 min.
Prereqs: Databricks AI/BI
Why this matters
Dashboards and Genie Spaces cover standard reporting and ad-hoc queries. Some use cases need a fully custom UI — an approval workflow, a data quality monitor, a customer-facing portal. Databricks Apps lets you deploy web applications (Streamlit, Dash, Gradio, or any framework) that run inside the Databricks environment with built-in authentication and direct access to Unity Catalog tables.
How it works
A Databricks App is a containerized web application deployed to a Databricks-managed endpoint. The app inherits the deployer's service principal credentials, so it can read and write tables without managing separate connection strings. Users authenticate through the workspace SSO — no extra login flow required.
Apps are defined by an app.yaml configuration file and deployed via the Databricks CLI or UI.
The Databricks Apps Cookbook has ready-to-deploy examples covering Streamlit, Dash, Gradio, and React patterns.
When to use / when not to
Use Databricks Apps when:
- You need a custom UI that goes beyond charts and tables (forms, workflows, interactive tools).
- The app needs direct, governed access to lakehouse data without ETL to an external database.
- You want single sign-on through the workspace — no separate auth infrastructure.
Use dashboards or Genie Spaces instead when:
- The use case is standard reporting or ad-hoc questions.
- No custom interactivity is required.
Common pitfalls
- Building an app when a dashboard would suffice — apps require development and maintenance. If the use case is "show charts with filters," a dashboard is faster and cheaper.
- Not using a service principal — apps should run under a service principal with the minimum grants needed. Running under a personal identity creates a single point of failure when that person leaves or rotates credentials.
Next
- Learn why: Databricks AI/BI overview
- Reference: Databricks Apps — Databricks docs
- Resource: Databricks Apps Cookbook