Google Launches A2UI 1.0: An Open Protocol for Generative UI and Agentic Experiences

Originally published on PkLavc.

Confirmed facts

Google announced A2UI 1.0 on 2026-11-11. The confirmed summary describes it as an open protocol for generative user interfaces and agentic experiences.

The protocol is described as allowing AI agents to render interactive, native screens across web, mobile, and desktop applications.

The release information lists open-source renderer SDKs for React, Angular, Flutter, and SwiftUI, and agent SDKs for Python, TypeScript, and Dart.

The confirmed key features include bidirectional function calls, swappable inference formats, and secure data binding.

• A2UI 1.0 was announced as an open protocol for generative UI.

• The protocol was described as enabling AI agents to render interactive native screens.

• Renderer and agent SDK support was announced.

• Bidirectional function calls, swappable inference formats, and secure data binding were announced.

Technical analysis

The technical analysis starts with the disclosed scope. The confirmed announcement describes a protocol, renderer SDKs, agent SDKs, and interaction features. The disclosed purpose is to let agents render interactive native screens. The disclosed renderer SDK list and agent SDK list establish a cross-framework integration surface for engineering teams.

The bidirectional function calls feature is described, but the approved evidence does not provide implementation requirements or runtime semantics. Swappable inference formats and secure data binding are described. The approved source did not disclose performance measurements, security threat models, or deployment architecture.

The evidence supports an inference about integration work: teams evaluating agentic interfaces would need to map agent actions to framework component APIs. That inference is tied to the disclosed SDK scope and is not a confirmed product capability.

• The disclosed SDK scope is an integration surface for agent and UI code.

• The confirmed interaction feature raises engineering questions about function-call handling.

• The approved source did not disclose implementation, training, performance, or deployment details.

Continue reading: Full technical article on PkLavc

Explore more technology articles: PkLavc Tech Blog

Telegram: https://t.me/PkLavcTech
Discord: https://discord.gg/zder5EZRK