AG-UI is an open, lightweight, event-based protocol that standardizes how AI agents connect to front-end applications.
➡️ See demos of AG-UI-enabled applications here
AG-UI is an open, lightweight, event-based protocol for agent-human interaction, designed simplicity & flexibility:
- During agent executions, agent backends emit events compatible with one of AG-UI's 16 standard event types
- Agent backends can accept one of a few simple AG-UI compatible inputs as arguments
AG-UI includes a flexible middleware layer that ensures compatibility across diverse environments:
- Works with any event transport (SSE, WebSockets, webhooks, etc.)
- Allows for loose event format matching, enabling broad agent and app interoperability
It also ships with a reference HTTP implementation and default connector to help teams get started fast.
AG-UI was developed based on real-world requirements and practical experience building in-app agent interactions.
It was shaped through:
- Working with users in the CopilotKit community to understand the needs of agent-based interactions in applications
- Collaborating closely with leading agent frameworks (LangGraph, Mastra, CrewAI, AG2, etc.)
- Extracting common infrastructure patterns that emerged across varied frameworks into a standardized, open protocol
This practical approach has helped ensure that AG-UI is both flexible and immediately applicable in real-world use cases.
Play with this hello-world app here: https://agui-demo.vercel.app/
Video:
agui-hello-world-2.mp4
AG-UI works with several popular agent frameworks and frontend solutions:
AG-UI-compatible agents securely and efficiently support:
- 💬 Agentic chat with real-time streaming
- 🔄 Bi-directional state sync (in and out of chat)
- 🧩 Generative UI and structured messages (with delta streaming)
- 🧠 Realtime context enrichment
- 🛠️ Frontend tool use (tool calls)
- 🧑💻 Human-in-the-loop and human-on-the-loop collaboration
And more. Learn more at ag-ui.com
The ag-ui dojo showcases many of the building blocks that ag-ui supports.
The building blocks are designed to be simple and focused -- between 50-200 lines of code.
dojo-hello.mp4
AG-UI provides ready-to-use client libraries for building connected experiences:
- React client via CopilotKit
- Messaging clients (WhatsApp, WeChat, RCS) — WIP in collaboration with AWS SNS
Learn how to build AG-UI hooks for a new framework: Framework Integration Guide
We 💜 contributions! Whether you're fixing bugs, improving documentation, or building demos — your contributions make AG-UI better.