"Create AI Apps in 5 Minutes with Databricks: A Step-by-Step Guide for Developers"

Databricks has simplified app development with the introduction of Databricks Apps, enabling enterprise developers to create production-ready data and AI applications in just a few clicks.

Now available in public preview, this service offers a template-driven experience that allows users to connect relevant data and selected frameworks to build fully functional apps within their Databricks environment.

According to the company, users can create and deploy secure applications in as little as five minutes.

This announcement arrives as enterprises, eager to leverage data-driven applications, face challenges managing the entire development cycle, from provisioning the infrastructure to ensuring security and access control.

Key Features of Databricks Apps

Like Snowflake, Databricks has long enabled users to build applications using data stored on its platform, including interactive dashboards and sophisticated AI systems like chatbots and fraud detection tools. However, the process of developing reliable, secure apps has often been complex and time-consuming.

Developers previously needed to grapple with various critical facets of the development pipeline—managing infrastructure, ensuring data governance, and setting up access controls. This made app production a daunting task.

Shanku Niyogi, VP of Product Management at Databricks, noted, “App authors had to navigate container hosting technologies, implement single sign-on authentication, configure service principals and OAuth, and manage networking—creating apps relied on integrations that were often brittle and hard to manage."

To address these issues, the new Databricks Apps experience consolidates all necessary functionalities into one platform.

Users simply select a Python framework (Streamlit, Dash, Gradio, or Flask), choose an app template (like a chatbot or data visualization app), and configure basic settings such as resource mapping and permissions. After this straightforward setup, the app is deployed within the user's Databricks environment, where it can be shared with team members. When they log in, the app automatically prompts for single sign-on authentication. Developers can even customize the app and test their code in their preferred integrated development environment (IDE).

On the backend, Databricks provisions serverless compute to run the applications, ensuring rapid deployment while keeping data secure within the Databricks environment.

Niyogi emphasized, "Each app is equipped with robust security measures for seamless user access. The integration with Unity Catalog ensures comprehensive data governance and management, inheriting the workspace's networking protections for a multi-layered security approach."

Expanding Framework Support

Currently, Databricks Apps supports Python frameworks. However, the company plans to introduce additional tools, languages, and frameworks to make secure app development accessible to a wider audience.

“We’ve started with Python, the leading language for data. Anyone familiar with a Python framework can easily write and onboard their app into Databricks Apps, which supports any Python IDE. We are collaborating with ISV partners to enhance compatibility and add support for other languages,” Niyogi stated.

Around 50 enterprises, including Addi, E.ON Digital Technology, SAE International, Plotly, and Posit, have already tested Databricks Apps in beta. With today’s public preview launch, this number is expected to grow.

Notably, Snowflake, Databricks' primary competitor, also offers a low-code solution for app development. However, Databricks asserts that its flexible and interoperable approach sets it apart.

Niyogi highlighted, "Databricks Apps supports Dash, Gradio, Flask, Shiny, and Streamlit, accommodating more Streamlit versions than Snowflake. Developers can utilize their preferred tools for app development, and we will continue enhancing this flexible approach with additional languages, frameworks, and tools."