Google has released A2UI 1.0, the first stable version of its open protocol for generative user interfaces. The technology gives AI agents a way to respond with interactive screens instead of sending users only written answers.
This could change how people use AI inside websites, mobile apps and business software. A user may ask an agent to compare products, prepare a plan or collect information. Instead of receiving a long message, the user could get a screen with charts, forms, choices and buttons.
What is A2UI?
A2UI stands for Agent-to-User Interface. It is a protocol that helps an AI agent describe the interface it wants to show.
The important part is that the AI agent does not need to draw the final screen itself. The app receiving the instructions uses its own registered components to display the interface.
For example, a travel assistant could create a flight selection screen with dates, prices and a confirmation button. A shopping assistant could show product cards and comparison charts. A business agent could ask a team member to approve a task using a small interactive card.
The app remains in control of its own visual design and components. This means developers can keep their existing brand style instead of allowing every AI agent to create a completely different interface.
What is new in version 1.0?
The stable release adds bidirectional function calls. This allows an interface to request an action and receive information back from the system. A flight or reservation picker, for example, could check live availability before asking the user to confirm.
A2UI 1.0 also adds support for accessibility and extra metadata. These features matter because an interactive screen must work for different users and fit the needs of the app that displays it.
Google has also expanded the developer tools around the protocol. Open-source renderers support React, Angular and Lit on the web, while Flutter can be used across mobile, desktop and web projects. SwiftUI is available for iOS, and Jetpack Compose support is being updated towards version 1.0.
Developers can also use agent SDKs for Python, TypeScript and Dart. These tools help connect the protocol to an existing AI agent or application framework.
Several companies are already using it
Google says companies such as Atlassian, Alation, Bloomberg and ByteDance are building with A2UI.
One example involves a work-management tool. An AI agent can ask a team member for a decision inside a task card. The user can provide the missing information without leaving the application or writing a long reply.
Another example involves business data. An agent can show charts and information cards that can be viewed inside an AI assistant or the company’s own application.
How does it work with AI assistants?
A2UI can also work with MCP Apps, a system that allows supported AI assistants to host interactive applications. Google says its web and Flutter renderers can be used in compatible MCP Apps, including environments that support assistants such as Gemini and Claude.
This does not mean every AI assistant will automatically support every A2UI feature. Developers still need to build the app, register the components and configure the connections.
Why does this matter for users?
Text is useful for asking questions, but it is not always the best way to complete a task. Comparing several phones, choosing a flight or checking a business report can be easier with a visual screen.
A2UI gives developers a common way to build these experiences across platforms. It could help AI assistants become more useful inside the apps people already use.
Also Read
- Google brings Gemini into Chrome on Android
- Gemini 4 Argon and advanced AI reasoning
- Gemini 4 Carbon and advanced AI workflows
- Amazon TorchNeuron for AI applications
- Google Pixel Watch 4 and its AI features
Frequently asked questions
Is A2UI an AI chatbot?
No. It is an open protocol that helps AI agents create interactive user interfaces inside supported applications.
Is A2UI 1.0 open source?
Google has released open-source renderers and SDKs for developers to build with the protocol.
Does A2UI work on mobile phones?
Yes. Developers can use Flutter and SwiftUI, while Android support is being updated through Jetpack Compose.
Passionategeekz Take
A2UI could become an important part of the next generation of AI apps. It gives developers a way to turn an AI answer into something people can actually use and interact with.
The biggest test will be adoption. If more apps and AI assistants support the protocol, users may start seeing fewer long AI replies and more useful screens that help them complete tasks.
[Via]