What problem does it solve?
Builds and deploys AI agents that can stream conversations and tool calls using the AG-UI protocol on CloudBase, reducing the effort of wiring frameworks, events, and UI clients by hand.
Core Features & Use Cases
- AG-UI Protocol Support for Streaming: Understand and implement the required event lifecycle (RUN_STARTED/RUN_FINISHED/RUN_ERROR plus TEXT_MESSAGE_* and TOOL_CALL_* patterns) for reliable UI updates.
- Production-Oriented Agent Adapters: Use ready adapters for LangGraph and LangChain, or create custom adapters by extending AbstractAgent and returning Observable<BaseEvent>.
- Agent Server Deployment Guidance: Deploy agent HTTP services using @cloudbase/agent-server with CloudBase best practices (Node.js 20 runtime, SSE readiness, dependency alignment, and manageAgent MCP).
- UI Client Integration: Connect web and WeChat Mini Program UIs to the agent server endpoints using @ag-ui/client and @cloudbase/agent-ui-miniprogram.
Quick Start
Ask your coding agent to deploy a TypeScript AG-UI agent server on CloudBase using manageAgent with Node.js 20, then wire a LangGraph or LangChain adapter and confirm SSE streaming works from the /send-message endpoint.