A2UI 1.0 SDKs for React, Flutter, SwiftUI and Jetpack Compose enable secure generative UI agents

Google published a blog post on its Developers site announcing the release of A2UI 1.0, describing it as the first stable version of an open protocol for generative user interfaces.reported The update allows artificial intelligence agents to render interactive, native screens across web, mobile and desktop applications instead of relying solely on text outputs.This update lets AI agents show interactive screens like forms and charts instead of just text.The release introduces bidirectional function calls and secure server-side data binding for native UI rendering.

The evidence

According to the Google Developers blog, the protocol enables agents to describe what to show while each application draws the interface using its own native components.reported This approach ensures that users experience low latency and complete native fidelity whether they interact with an agent inside a company's own app or within a third-party assistant such as Gemini or Claude.company claim The release includes open-source renderer SDKs for React, Angular, Lit, Flutter, SwiftUI and Jetpack Compose, alongside agent SDKs for Python, TypeScript and Dart.

The decision

The blog post states that companies including Atlassian, Alation, Bloomberg and ByteDance are already building with the protocol. By adopting A2UI, developers can leverage bidirectional function calls and swappable inference formats to integrate generative UI into their product stacks.reported The system allows agents to use a company's specific design system, placing and configuring only the components that the developer has registered, which maintains brand consistency.

What to make of it

The source material highlights that current user interactions with agents resemble the terminal era, where capable systems answer almost everything in text. This often results in walls of text that make it difficult to compare products or edit detailed plans. A2UI aims to solve this by allowing agents to respond with interactive elements such as forms to confirm actions or charts to display data, adapting to each user's unique needs in the moment.reported

In practice

For startup founders, the immediate change is access to a stable, versioned protocol for building agentic interfaces. The documentation confirms that developers can use existing open-source SDKs for major frameworks like React and SwiftUI to render these interfaces securely. However, the protocol requires that applications render content at native speed and handle data binding on the server to ensure sensitive values never reach the language model. Availability is confirmed for web, Android and iOS platforms, with agent support for Python, TypeScript and Dart. No specific pricing or licensing fees are mentioned in the source material, but the SDKs are described as open-source.