Google has announced the stable release of A2UI 1.0, an open protocol for generative UI. The protocol enables AI agents to render interactive native interfaces instead of simply returning text, and introduces features such as bidirectional function calling and accessibility support. Its security design ensures that sensitive values never enter the model. The protocol is open-sourced under the Apache 2.0 license, was created by Google, and is jointly maintained by the relevant team and the open-source community. The official release also ships with open-source renderers covering multiple platforms, new TypeScript and Dart agent SDKs, and support for enterprises to register A2UI agents in Gemini Enterprise. Companies including Atlassian, Alation, and ByteDance have already implemented the protocol, and Google's own products such as Gemini Notebook have adopted it as well. Documentation and quick-start guides are available at a2ui.org.