AGenUI/AGenUI

What it solves

AGenUI solves the challenge of rendering interactive, LLM-generated user interfaces (Generative UI) on mobile devices in real time. It allows developers to build native experiences where the UI is not static but is streamed and updated dynamically based on the structured data emitted by a Large Language Model.

How it works

AGenUI implements the A2UI v0.9 protocol (open-sourced by Google). It uses a cross-platform shared C++ core engine to handle streaming data parsing, virtual component tree management, and style resolution. This core engine then dispatches component protocols to platform-specific rendering engines for iOS, Android, and HarmonyOS, which use native system capabilities to draw the UI components.

Who it’s for

It is designed for mobile app developers building AI-powered applications that require dynamic, native-feeling interactive interfaces (such as cards, forms, and lists) generated on-the-fly by an AI agent.

Highlights

• Multi-platform Support: Native rendering for iOS, Android, and HarmonyOS.
• Real-time Streaming: Renders structured component descriptions incrementally as they arrive from the LLM.
• Extensive Component Library: Includes 18 A2UI protocol components and 4 SDK extension components (e.g., Table, Carousel, Web).
• Customizability: Provides an API to register custom native components and integrate client-side tools via Function Call integration.
• Agent Integration: Includes a standalone A2UI generation Skill that can be mounted into AI coding agents (like Claude Code or Cursor) to guide them in generating conformant UI protocol messages.