Native SDK is an open-source toolkit designed for building native desktop applications that prioritize both performance and expressive UI authoring. Created to bridge the gap between web-based development models and native OS execution, the toolkit replaces heavy web runtimes with a native engine. It allows developers to write interfaces using declarative markup and Zig, which the SDK engine then renders directly into real operating system windows. This approach avoids browsers and WebViews, ensuring high performance, low memory usage, and authentic native experiences like OS-native scroll physics and context menus.

The framework follows a predictable, message-based state model where events trigger messages, messages update the state, and the interface re-renders based on that state. This architecture is designed to be highly readable, maintainable, and predictable, making it well-suited for collaboration between human developers and AI agents. The toolkit is cross-platform, natively supporting macOS, Linux, and Windows, with experimental support for iOS and Android platforms riding the same underlying runtime. It emphasizes high-quality defaults, including a built-in component catalog featuring typography, spacing, and color design tokens, while remaining customizable to accommodate specific app identities.

Some of the key features are:

• Native Rendering Engine: Renders widgets as actual OS windows without browsers or WebViews, achieving high-performance results with low resource consumption.
• Declarative Markup: Uses a dedicated markup language for defining UIs, complemented by Zig for business logic and state management.
• Predictable State Loop: Implements a strict model-view-update (MVU) pattern that ensures UI consistency and simplifies debugging.
• Component Library: Includes a comprehensive catalog of built-in components like buttons, charts, accordions, and virtualized lists that are fully styleable.
• AI Agent Integration: Provides native automation tools, allowing AI agents to drive the application, verify state, and generate UI reliably through a standardized automation server.
• Single Binary Releases: Packages the entire application, including the engine and assets, into a single, small, and fast-starting executable file.
• Native Platform Integration: Supports deep system-level integration, including native system trays, keyboard shortcuts, context menus, and file system dialogs.

The tool operates through a CLI that manages project scaffolding, hot-reloading development servers, and packaging. During development, the engine watches markup files and updates the running window in real-time without losing application state. For release, the toolkit compiles the markup, resulting in optimized native binaries that carry no interpreter or scripting engine. The build process is streamlined, and applications are packaged as distributable bundles for each target OS, ensuring that end-users enjoy a native, installation-friendly experience.

Some common use cases include:

• High-Performance Dashboards: Building responsive data dashboards that require low latency and smooth interactivity using the native renderer.
• Native CLI Tools with UI: Creating desktop-based interfaces for complex command-line utilities to provide users with a graphical front-end.
• AI-Assisted Applications: Developing applications where AI agents are integrated into the workflow to assist with code generation and testing.
• Cross-Platform Utilities: Maintaining a single codebase that delivers high-fidelity desktop applications across Windows, macOS, and Linux.