Watch: A2UI (Agent-to-User Interface)
An open web standard that lets AI agents generate rich, interactive interfaces as declarative data, rendered through a client app's own trusted components instead of raw code.
Transcript
When an AI agent needs to show you a form, a chart, or an interactive widget, how does it do so securely? That is where a new open web standard called Agent-to-User Interface, or A2UI, comes in.
Traditionally, having an AI generate custom user interfaces can be a security risk if the agent tries to run raw, executable code on your device. A2UI solves this by allowing the agent to send rich, interactive elements as safe, declarative data instead of code.
The agent simply describes what the interface should contain, such as a questionnaire or a graph. Then, the host application takes that data and renders it using its own trusted, pre-built components. This keeps a clear security boundary in place, dramatically reducing the risk of malicious injections or unsafe rendering.
A2UI is designed to work alongside other emerging agent standards, like the Model Context Protocol and the Agent Payments Protocol. Whenever an AI assistant needs to hand something interactive back to a human, A2UI ensures the experience is both seamless and secure.