What is a data app?

Summary

• A data app is an interactive application that puts data, analytics, and AI capabilities directly in the hands of users through a web-based interface, so business and technical users can act on data without moving it out of your platform.
• Databricks Apps lets you build and deploy data apps as secure data and AI applications that run directly on the Databricks platform, with no separate infrastructure to provision or manage.
• The application comes to the data. Apps run next to your governed data instead of copying it to an external system, which reduces data movement, latency, and operational overhead.
• Build with familiar tools. Write apps in Python (Streamlit, Dash, Gradio, Flask) or Node.js (React, Angular, Svelte, Express) in IDEs like Visual Studio Code, then deploy to a workspace.
• Governed and serverless by default. Apps use built-in authentication, enforce Unity Catalog permissions, and run on fully managed serverless compute that scales automatically.

What is a data app?

A data app is an interactive application that delivers data, analytics, and AI capabilities to users through a web-based interface, letting them explore data, run workflows, and take action directly. Databricks Apps is the platform capability for building and deploying these applications: it enables developers to build and deploy secure data and AI applications directly on the Databricks platform, without provisioning or managing separate infrastructure. The core idea is that the application comes to the data, rather than moving data out to another system. See Databricks Apps documentation.

Why build data apps on the Databricks Platform

• Build with familiar frameworks and tools. You can develop apps locally in Python or Node.js and deploy them to a workspace. Supported Python frameworks include Streamlit, Dash, Gradio, and Flask, and supported Node.js frameworks include React, Angular, Svelte, and Express. You can develop in familiar IDEs such as Visual Studio Code. See Introducing Databricks Apps.
• Fully managed serverless hosting. Each app runs as a containerized service on Databricks serverless compute, with its own configuration, identity, and isolated runtime environment. Compute is provisioned automatically and scales for you, so there is no infrastructure to manage, and apps are billed per hour of compute while running. See App key concepts.
• Governed by Unity Catalog. Apps integrate with Unity Catalog for granular access control and data lineage, and they enforce user-level access based on Unity Catalog permissions. Data leaves your Databricks environment only if you choose to share it.
• Secure by default. Apps provide out-of-the-box authentication using OAuth 2.0 / OIDC and single sign-on, use automatically managed service principals for application-to-application communication, and include built-in audit logging to track usage and access.
• Fits your development practices. Apps support Git version control and CI/CD pipelines, so they fit into existing software development workflows. See Announcing the general availability of Databricks Apps.

Getting started

1. Choose a framework you already know in Python or Node.js and scaffold your app.
2. Develop and test locally in an IDE such as Visual Studio Code, connecting to your governed data through Unity Catalog.
3. Deploy the app to a Databricks workspace, where it runs on serverless compute with built-in authentication and Unity Catalog enforcement.
4. Follow the step-by-step guidance in the Databricks Apps documentation.

FAQs

What can you build with a data app?

Common data apps include interactive dashboards, retrieval-augmented generation (RAG) chat applications, data entry and labeling forms, custom operational interfaces, and AI agents that combine tools, large language models, and vector search.

Do you need to manage servers to run a data app?

No. Databricks Apps provides fully managed serverless hosting: compute is provisioned automatically and scales for you, and there is no infrastructure to manage.

How is data in a data app governed?

Apps run next to your data on the platform and enforce user-level access based on Unity Catalog permissions, with data lineage and built-in audit logging. Data leaves your environment only if you choose to share it.