Google Ships A2UI 1.0 Letting Agents Render Native UI
Google released A2UI 1.0, the first stable version of its open generative-UI protocol, in which agents describe what to display while each app renders it with its own native components, preserving low latency and native fidelity across web, Android, iOS and MCP hosts. The release adds bidirectional function calling and accessibility/metadata extensions, shipping renderers for React, Angular, Lit, Flutter, SwiftUI and Jetpack Compose plus Python, TypeScript and Dart agent SDKs; Atlassian, Bloomberg and ByteDance are already building on it.
Related Entities
Full Report
IT之家 November 12 news, A2UI is an open protocol for generative UI announced by Google at the end of 2025, allowing agents to not only reply to users with text but also directly reply with UI interface content.
Today, Google officially released A2UI 1.0, the first stable version of the protocol. Developers can now use it together with open-source web, Android, and iOS renderer SDKs, as well as agent SDKs for Python, TypeScript, and Dart. Companies such as Atlassian, Alation, Bloomberg, and ByteDance have already begun developing with it.
Google explains that in A2UI, the agent is responsible for describing what content to display, and each app can draw the UI using its own native components. Whether users access the agent within a company's own app or through a third-party AI assistant hosting an MCP app (such as Gemini or Claude), they get the same low latency, high trust, and full native fidelity.
▲ The same request answered via text and A2UI, then the same view on a laptop
IT之家 learned from Google's official blog that A2UI 1.0 adds bidirectional function calling, as well as extended support for accessibility and additional metadata.
The open-source renderers can draw A2UI screens on the web using React, Angular, and Lit; render on mobile, web, and desktop with Flutter; and use SwiftUI on iOS and Jetpack Compose on Android.
The web and Flutter renderers also run inside MCP apps, so A2UI content can appear in AI assistants that support MCP apps, such as Claude and Gemini.