Agentic AI
Google Releases A2UI 1.0, Bringing Interactive Screens to AI Agents
A new open protocol lets AI agents respond with interactive components instead of plain text, enabling richer user experiences across web and mobile platforms.
Google Developers Blog (AI)
Google has released A2UI 1.0, an open protocol that enables AI agents to communicate through interactive user interfaces rather than walls of text. The framework allows developers to build agents that respond with forms, charts, and other native UI components tailored to each platform.
What Is A2UI?#
A2UI (Agent-to-UI) bridges the gap between what agents can accomplish and how users experience those accomplishments. Instead of an agent returning lengthy text responses to requests, it can construct adaptive interfaces that display information visually and allow users to interact directly with options. Each application renders the interface using its own native design system and components, ensuring consistency with brand guidelines while maintaining high performance.
The protocol works by having the agent describe what should appear on screen, and then each platform's renderer displays it natively on web, Android, or iOS. This approach keeps sensitive data on the server—the language model never handles user credentials or private information.
Key Features of Version 1.0#
The stable release introduces several capabilities for production deployment:
- Bidirectional function calls enable interactive components like flight or reservation pickers to check availability in real time
- Accessibility support and extensions for additional metadata
- Open-source renderer libraries for React, Angular, and Lit on the web; Flutter across platforms; SwiftUI on iOS; and Jetpack Compose on Android
- Agent SDKs for TypeScript, Dart, and Python that streamline integration and reduce latency through swappable inference formats
Real-World Applications#
Companies including Atlassian, Alation, Bloomberg, and ByteDance are already building with the protocol.
Atlassian is integrating A2UI into Jira, allowing agents to ask teams for follow-ups—such as clarifications or decisions—within work items themselves rather than in a separate chat interface. Alation uses A2UI to render its Intelligence Operating System feeds with metric cards and charts, which display consistently whether viewed in Claude or in Alation's native application. built an A2UI renderer for Lynx, its cross-platform framework, and demonstrated it with a trip-planning interface. 's own Gemini Notebook uses A2UI cards to let users refine preferences—such as slide deck format and presentation style—before generation begins.

