Amazon Bedrock AgentCore gets AG-UI protocol for interactive agent frontends
New open protocol lets AI agents render inline charts and wait for human approval.
AG-UI (Agent-User Interaction Protocol) is an open standard that defines how agent backends communicate dynamic events to frontends, enabling generative UI, shared state, and human-in-the-loop workflows. It works with multiple agent frameworks (Strands Agents, LangGraph, CrewAI) and frontend libraries (React, Angular, Vue), keeping backend and frontend code decoupled. Amazon Bedrock AgentCore now supports the AG-UI protocol, acting as a transparent proxy that handles authentication (SigV4 or OAuth 2.0 via Cognito), session isolation, scaling, and observability while passing AG-UI requests to your container on port 8080. The Fullstack Agentcore Solution Template (FAST) v0.4.1 ships two AG-UI agent patterns (agui-strands-agent and agui-langgraph-agent) with a shared frontend parser that works regardless of the backend framework.
CopilotKit extends FAST with generative UI (inline charts and components), bidirectional shared state (a todo canvas), and human-in-the-loop interactions (a meeting scheduler that pauses the agent for user input). Both layers deploy on AgentCore Runtime with Cognito authentication and AgentCore Gateway for MCP tool connectivity. This integration allows developers to build interactive, real-time agent frontends without vendor lock-in, using any framework and model. The solution is available via the FAST repository and CopilotKit sample on GitHub, with full deployment instructions using AWS CDK, Amazon ECR, and AWS Amplify.
- AG-UI is an open protocol that standardizes generative UI, shared state, and human-in-the-loop interactions between agents and users.
- Amazon Bedrock AgentCore supports AG-UI as a transparent proxy, handling auth (SigV4/OAuth), scaling, and session isolation.
- FAST v0.4.1 includes AG-UI patterns for Strands Agents and LangGraph, with CopilotKit adding inline charts, shared canvases, and approval workflows.
Why It Matters
This standardizes interactive agent UIs, enabling richer user experiences and faster cross-framework development for enterprise AI applications.