Build delightful, fast, reliable
agentic experiences.
A2UI is an open-source generative UI framework that enables agents to select, compose, or build experiences that are personalized in the moment, native to your brand, and reliable and secure. Using an agent-optimized protocol, native renderers, and SDKs for agent frameworks and harnesses, A2UI achieves low latency, predictability, and visual fidelity across all screens.
Custom UI for each user's intent.
Dinner for four tonight, one vegetarian, under $40.
Personalized in the moment
Your agent constructs UI that adapts to each user's unique needs and preferences.
Native to your brand
Your design system, locked. Agents place only components you registered.
Fast, reliable, secure in production
Agents compose UI with A2UI Express or Jev-style decision models. Macros bind data and components reliably, and the client streams it in as the agent thinks.
Generative UI spectrum with A2UI
The same request can come back as a pre-built widget, a composed screen, or generated code. Move the slider to see each one and how A2UI handles it.
“Flight delayed — rebook me on the next nonstop with an aisle seat.”
FlightCard + ChoiceChips + Button Tailored layouts assembled from your native catalog.
- Adapts to every prompt Agents compose multi-step flows on the fly from widgets you approve.
- Native & safe by default Streams declarative JSON data with two-way bindings — never executable code.
- One stream, every screen Renders natively across React, Lit, Angular, Flutter, SwiftUI, and Compose.
With A2UI: The model builds the screen from components in your catalog, and the renderer draws it with your own components as it streams in. In hosts that support MCP Apps, the same A2UI surface can run inside an MCP App.