A2UI is an open-source protocol developed by Google that enables AI agents to safely send declarative component descriptions for rendering rich user interfaces across trust boundaries. Instead of relying on text-only responses or risky code execution, agents use a flat, streaming JSON structure to describe UIs that clients render using their own native widgets. This framework-agnostic approach allows the same agent response to work seamlessly across Angular, Flutter, React, and native mobile applications with custom-styled components.
Key features include:
- Security: Declarative data format prevents UI injection attacks by limiting agents to pre-approved components.
- LLM-Friendly: Designed for easy generation with incremental UI building capabilities.
- Progressive Rendering: Streams UI updates in real-time as they are generated.
- Cross-Platform Compatibility: Renders natively on web, mobile, and desktop environments.
Use cases span AI-driven applications where agents need to generate interactive interfaces, such as landscape architect tools, data visualization dashboards, and custom form builders, all while maintaining security and performance.
