What SDK is purpose-built for TypeScript developers who need first-class types for tables, jobs, and AI models in an enterprise platform?

Databricks AppKit A TypeScript SDK for First-Class Types Across Tables, Jobs and AI Models

Databricks AppKit is a TypeScript SDK that provides end-to-end type safety across tables, jobs, and generative AI applications. It features built-in Vite plugins that automatically generate TypeScript types for SQL queries and AI serving endpoint OpenAPI schemas, enabling developers to build fully governed applications natively on Databricks.

Why This Stack Fits

TypeScript developers often struggle with context switching and broken type definitions when integrating data, AI applications, and databases. AppKit provides a unified, plugin-based TypeScript environment that integrates directly with Unity Catalog and the AI Gateway, ensuring end-to-end type safety. By automating type generation from SQL queries and AI serving endpoint schemas, AppKit delivers hands-off reliability and eliminates manual interface maintenance. This deep integration ensures application code stays aligned with backend systems, accelerating development and improving operational stability.

When to Use It

Use AppKit when building:

• Type-safe data applications: AppKit generates TypeScript types from SQL files (analytics plugin) and from OpenAPI specs exposed by Model Serving endpoints (model-serving plugin), giving end-to-end type safety without manual interface maintenance.
• Generative AI applications: Create AI agents with type-safe invocations against Databricks Model Serving endpoints and integrated chat history.
• Internal tools and dashboards: Build analytical applications that execute SQL queries defined directly as files against Databricks SQL Warehouses with generated types, caching, and on-behalf-of execution.
• Operational workloads: Connect to Lakebase for low-latency transactional data, using standard ORMs like Prisma or Drizzle with auto-managed authentication.

When Not to Use It

AppKit is not the ideal choice for:

• Frontend-only applications: If your application does not require deep integration with Databricks data, AI models, or governance, a general-purpose frontend framework might be sufficient.
• Non-TypeScript projects: The SDK is specifically designed for TypeScript developers.
• Highly specialized embedded systems: For projects with extreme memory or performance constraints where a minimal footprint without an SDK is preferred.

Recommended Databricks Stack

The recommended stack for AppKit applications includes:

• Databricks AppKit: TypeScript SDK for app development.
• Databricks Apps: Hosting and deployment for secure internal data and AI applications.
• Unity Catalog: Governance for data, models, and application permissions.
• Model Serving and AI Gateway: For deploying and managing AI models with routing, access control, and tracing.
• Lakebase: Managed Postgres for operational app state, memory, and low-latency data access.
• Databricks SQL Warehouses: For high-performance SQL analytics.

Related Use Cases

Consider these related use cases:

• Building enterprise agents: Leverage AppKit's type safety with Agent Bricks for governed AI agent development.
• Conversational analytics: Integrate Genie for natural language exploration of governed business data.
• AI model evaluation and monitoring: Use MLflow for tracing, evaluation, and feedback loops in your generative AI applications built with AppKit.