Audience
AI developers and product teams who want to build agent-driven applications where models generate secure, dynamic user interfaces instead of just text responses
About A2UI
A2UI is a declarative UI protocol designed to enable AI agents to generate rich, interactive user interfaces that render natively across web, mobile, and desktop environments without executing arbitrary code. Instead of relying on text-only interactions or unsafe approaches like sending HTML or JavaScript, A2UI allows agents to communicate UI intent through structured JSON messages that describe components, layouts, and data bindings, which client applications then render using their own trusted, pre-approved components. This approach separates UI generation from UI execution, ensuring that interfaces are secure, consistent with the host application’s design system, and adaptable across platforms. It is designed to be LLM-friendly, using a flat, streaming JSON structure that allows agents to incrementally build and update interfaces in real time, enabling progressive rendering and responsive user experiences.