# Core principles

Learn about the fundamental concepts and principles behind AppKit.

## 1. Highly opinionated[​](#1-highly-opinionated "Direct link to 1. Highly opinionated")

AppKit must provide a clear path with best practices for building Databricks applications. We provide strong defaults, with advanced customization when needed.

## 2. Built for application use cases[​](#2-built-for-application-use-cases "Direct link to 2. Built for application use cases")

This SDK is for application development, not infrastructure management. Databricks' internal implementation details must be abstracted. We're building an application SDK, not a service wrapper.

## 3. Delightful developer experience[​](#3-delightful-developer-experience "Direct link to 3. Delightful developer experience")

Every interface, doc, example, tool, and implementation must provide developer joy. Combined with the Highly Opinionated principle, this creates a true plug-and-play experience.

## 4. Zero-trust security[​](#4-zero-trust-security "Direct link to 4. Zero-trust security")

Minimize exposed surface area, fail safely by default, and validate all inputs. AppKit must always have a zero-trust mindset.

## 5. Optimized for humans and AI[​](#5-optimized-for-humans-and-ai "Direct link to 5. Optimized for humans and AI")

Developers and LLMs both use this SDK. Every API must be discoverable, self-documenting, and inferable by both types of users. Test with both.

## 6. Production-ready from day one[​](#6-production-ready-from-day-one "Direct link to 6. Production-ready from day one")

Even the smallest feature can be used by enterprise users, so everything shipped must be production-ready. Observability, reliability, and scalability since day one.

## 7. Layered extensibility[​](#7-layered-extensibility "Direct link to 7. Layered extensibility")

AppKit provides high-level plugins, low-level primitives, and extension points for custom plugins. It integrates into any application architecture and never blocks your path forward.