A2UI 1.0 稳定版发布，智能体用交互界面答复

Google released A2UI 1.0, the first stable version of its open protocol for generative UI, which lets AI agents answer with an interactive screen such as a form to confirm or a chart. The agent only describes what to show, and each app draws it with its own native components, limited to the ones the developer registered. Data is bound on the server, so sensitive values never reach the model. The web and Flutter renderers also work inside MCP Apps, so the same screen can appear in AI assistants that support them, such as Claude and Gemini. New in 1.0 are function calls in both directions, so a reservation picker can check availability on the spot. Atlassian already uses it in Jira, where an agent that needs a missing detail asks with a card inside the work item and the team answers there. Not everything is on 1.0 yet, the Android Compose renderer and Gemini itself still support 0.9.1. And the screens do not have to come from an LLM, since Google names decision models that pick components as the second way to compose them.