CopilotKit v1.50 is coming soon!

AG-UI Protocol

AG-UI Protocol

The Open Protocol for Agent<->User Interaction

AG-UI (Agent–User Interaction) is the general-purpose, bi-directional connection between a user-facing application and any agentic backend. It is the connective layer that transforms agents from background processes into true collaborators — transparent, reliable, and always aligned with the user.

AG-UI, Frameworks and LLMs

AG-UI isn't a siloed solution — it's being developed by the CopilotKit team in collaboration with leading frameworks and platforms developers already trust:

pydanticlang-graphmastraagno-aicrewaiagz-aillama-index

What AG-UI Enables

Shared State
end
start
Tool-Based GenUI
end
start
Subgraphs
end
start
Agentic Chat
end
start
Human in the Loop
end
start
Agentic GenUI
end
start
Predictive Updates

Shared State: Bi-directional synchronization of agent and application state — either read/write or read-only.

Why AG-UI Matters

For Developers

Build agentic apps faster with a standard interface for agent ↔ UI communication.

benefit-card

For Enterprises

Deploy with confidence, with governance and analytics built in at the boundary.

benefit-card

For the Ecosystem

Prevent fragmentation with a shared protocol designed for broad adoption.

benefit-card

Bringing AG-UI Features into your app with CopilotKit

CopilotKit builds on AG-UI with tools for production-ready agentic apps

CopilotKit AG-UI Clients

First-party React and Angular AG-UI clients, with community clients emerging across Golang, Rust, Java, and more.

  // React AG-UI Client
  import { useAGUI } from '@copilotkit/react'

  const MyComponent = () => {
    const { agent, ui } = useAGUI()
    // Your component logic
  }

Copilot Cloud - Hosted AG-UI Platform

Protection

Guardrails at the boundary: prevent prompt injection, block sensitive data leaks, enforce brand and compliance guidelines.

Cockpit

Coming Soon

AI product analytics for visibility into every agent ↔ user interaction.

Learning

Coming Soon

Support for self-improving agents through implicit human feedback (RLiHF).

Frequently Asked Questions

How does AG-UI compare to MCP or A2A?

While MCP (Model Context Protocol) and A2A (Agent-to-Agent) handle context and agent coordination, AG-UI defines the layer of interaction between the user, the application, and the agent.

A2UI? AG-UI? What's the difference?

Despite the similar acronyms, these are different and work well together. A2UI is a generative UI specification (from Google) - which agents can use to deliver UI widgets, where AG-UI provides the complete bi-directional runtime connection between any agentic backend and a user-facing application.

Why is AG-UI important for enterprise adoption?

It provides transparency, safety, and control at the most critical boundary — where users interact with agents.

Protocol Comparison
Agent ↔ User Interaction
Agent ↔ User Interaction

User, application, and agent synchronization

checkcheck
dividerfocus
Model Context Protocol
Model Context Protocol

Handles context and model communication

checkcheck
dividerfocus
Agent-to-Agent
Agent-to-Agent

Agent coordination and communication

checkcheck
dividerfocus
Are you ready?

Ready to Get Started?

Explore the protocol specification and start building agentic applications with CopilotKit and AG-UI.