Google released A2UI 1.0, which it describes as the first stable version of its open protocol for generative UI. The protocol lets AI agents reply with interactive screens, such as a form to confirm or a chart, instead of text. Google says text is a poor way to compare three products or approve a detailed plan.
The agent describes what to show, and each app draws it with its own native components. Google says agents place only the components a company registers from its design system, and that A2UI binds data on the server so sensitive values never reach the model. Open-source renderers cover React, Angular, Lit, Flutter and SwiftUI, with the Android Jetpack Compose renderer still on v0.9.1. Agent SDKs are available for Python, TypeScript and Dart. Version 1.0 adds bidirectional function calls, so a reservation picker can check availability on the spot, plus accessibility support.
The web and Flutter renderers work inside MCP Apps, so A2UI screens can appear in assistants such as Claude and Gemini. Organizations can register agents that respond with A2UI in Gemini, which supports version 0.9.1 today, and Google says it is upgrading Gemini Enterprise to 1.0. Google names Atlassian, Alation, Bloomberg and ByteDance as building with the protocol; ByteDance released a renderer for its Lynx framework in June 2026.