AI Dev 26 X SF | Atai Barkai: Fullstack Agents & Generative UI with AG UI

Why It Matters

Standardizing agent-to-UI integration and providing agent-native developer tools could accelerate production-grade, user-friendly AI experiences across enterprises and platforms, reducing implementation complexity and enabling broader adoption. That shift may fundamentally change how software interfaces are designed and monetized as AI mediates more human-technology interactions.

Summary

Atai Barkai of Copagit outlined the company’s developer infrastructure for user-interactive agents and introduced the Agent-User Interaction (AG UI/AGI) protocol, which standardizes how agent backends connect to user-facing interfaces. He argued that agentic systems break the traditional request-response web model—requiring streaming, reconnections, mixed structured/unstructured data and mid-run steering—and described Copagit’s open-source SDKs and cloud optimized for those needs. Barkai mapped the emerging “generative UI” landscape onto a spectrum from controlled to declarative to open-ended interfaces, demoing controlled components that let agents render deterministic, pixel-perfect elements like charts. He said the industry is moving from early text-based AI interfaces to richer, Mac/Windows-style AI UIs and highlighted broad adoption of the protocol by major cloud providers and enterprise customers.

Original Description

All UI will be AI. In this talk, Atai Barkai from CopilotKit explored the emerging space of Generative UI, and went over the practical building blocks emerging in the space.