Apps & dev tools
Enterprise low-code app platforms and Databricks Apps both let you build applications quickly, but they sit in very different places in your architecture: low-code platforms are general-purpose app platforms for business workflows, while Databricks Apps are data/AI-native apps that run directly on the Lakehouse. In practice they often complement each other with low-code tools integrating with Databricks with a standard set of patterns and Databricks Apps used for data-centric experiences.
Underlying principles
- Enterprise low-code platform: Optimized for line-of-business apps, workflow automation, portals, and integration across many systems (CRM, ERP, legacy, data platforms, etc.).
- Databricks Apps: Optimized for data processing & AI applications using Databricks-native features. Use Databricks SQL for queries and datasets, Lakeflow Jobs for batch processing, and Model Serving for AI inference workloads. Offload heavy data processing to these services to avoid performance issues.
Platform architecture
- Low-code platform architecture: Multi-tenant app runtime, visual builder, integration layer to many backends, with enterprise features (SSO, RBAC, environments, governance) as shared services.
- Databricks Apps architecture: Apps are hosted on Databricks' serverless platform, tightly integrated with Unity Catalog, compute clusters/serverless, Lakebase (OLTP on Lakehouse), and data/AI governance.
Data and integration model
- Low-code: Treats Databricks (and other systems) as data sources via JDBC/ODBC/REST, plus other SaaS and on-prem systems; often acts as orchestration layer for processes spanning multiple backends.
- Databricks Apps: Directly read/write Lakehouse and Lakebase tables, call ML models and streaming pipelines in the same environment; external integration usually via APIs or connectors from the data platform outward.
Development and users
- Low-code: Designed for citizen developers, Business Analyst and Full-Stack developers teams with visual modelling plus optional custom code; aims to standardize app delivery across many departments and often require a high degree of UI customization and managing transactional data.
- Databricks Apps: Targeted mainly at data engineers/scientists and Machine Learning engineers familiar with Python or Node.js.
Summary table
| Feature | Enterprise Low-Code Platform | Databricks Apps |
|---|---|---|
| Primary Goal | Accelerate full-stack application development (UI, business logic, transactions, workflows). | Accelerate deployment of data, ML, and AI applications (dashboards, inference, RAG UIs). |
| Development Method | Primarily visual modeling (drag-and-drop), BPMN diagrams, and proprietary microflows. | Primarily code-first (Python/Node.js) using open-source frameworks (Streamlit, Flask, Dash). |
| Deployment Model | PaaS-based, proprietary runtime engine, often utilizing Kubernetes/containers managed by the vendor. | Serverless compute managed by Databricks, running containerized code within the customer's VPC/cloud account. |
| Data Access | Relies on pre-built connectors (REST, SOAP, SAP) to external databases. | Native access to Unity Catalog data plus Model Serving, Databricks SQL, Vector Search, and Lakebase. |
| Data Governance | Centralized RBAC for the application layer, often custom-built. | Inherits Unity Catalog governance (fine-grained RLS/CLM) and workspace network security. |
| Scalability | High horizontal scalability, often configured manually via app tiers and environments. | Apps run on compute resources that determine processing power and memory. |
For many enterprises the target state is: core data and AI on Databricks, with Databricks Apps for data-heavy experiences, and a low-code platform for broader business workflows that call Databricks via APIs or SQL warehouses when needed.
Partner hosted apps
Partner-hosted apps are applications run and operated by a Databricks technology partner in that partner's own environment, securely connected to a Databricks Data Intelligence Platform. Databricks provides the data, governance, and workspace resources (e.g., SQL warehouse), while the partner is responsible for hosting, operating, and updating the application.
See integration requirements for the minimum technical requirements.
Databricks Apps
Databricks Apps are web applications that run as containerized services on the Databricks serverless platform. They're optimized for data-centric experiences that require tight integration with Unity Catalog and Databricks services.
For detailed architecture, development lifecycle, and implementation patterns, see Databricks Apps in Databricks Design Patterns.
What's next
- Review the integration requirements for foundational guidance
- Learn about telemetry and attribution for usage tracking
- Explore other Partner product categories for additional integration patterns
- See the Databricks Apps documentation for detailed implementation guidance