cloudbase-agent

Build AG-UI compatible TypeScript agent servers streaming runs to UI clients.

2|1|Updated Apr 3, 2026
One-click install
npx skills add https://github.com/lxb031018/qintu --skill cloudbase-agent-lxb031018
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: cloudbase-agent
Source: https://github.com/lxb031018/qintu/tree/main/rules/cloudbase-agent/ts
Command: npx skills add https://github.com/lxb031018/qintu --skill cloudbase-agent-lxb031018

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Cloudbase Agent (TypeScript) helps teams build and deploy AI agents that communicate with UIs through the AG-UI event protocol, so you can stream messages, tool calls, and state updates instead of building one-off HTTP handlers.

Core Features & Use Cases

  • Deploy AG-UI compatible agent servers: Expose HTTP endpoints (including SSE streaming) for agent runs and UI integration.
  • Adapter-based framework integration: Use LangGraph or LangChain adapters, or implement custom adapters by extending AbstractAgent.
  • Protocol-level control and interoperability: Emit and handle AG-UI lifecycle events (RUN/TEXT/TOOL/STATE) and support client-defined tools.

Quick Start

Ask the AI assistant to “Create a CloudBase AG-UI agent server using @cloudbase/agent-server with the LangGraph adapter, wire the server to stream events over SSE, and explain how the client should call the /send-message endpoint using threadId and runId.”

Frequently Asked Questions about cloudbase-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I stream LangGraph AI agent runs to a UI client using SSE?

To stream LangGraph agent runs over SSE, use the @cloudbase/agent-server with the LangGraph adapter, which handles emitting AG-UI protocol lifecycle events like RUN_STARTED, text streaming, and tool calls to connected UI clients.

What is the AG-UI event protocol lifecycle for TypeScript AI agents?

The AG-UI event protocol lifecycle for TypeScript AI agents requires emitting RUN_STARTED and RUN_FINISHED or RUN_ERROR events, alongside text streaming, tool call/result, and optional state snapshots to manage UI synchronization.

Can I build a custom AI agent adapter if I don't use LangChain or LangGraph?

You can build custom AI agent adapters by extending the AbstractAgent class to implement proprietary logic while still adhering to the AG-UI protocol for UI streaming, state snapshots, and tool calling.

Does the AG-UI client work with WeChat Mini Programs?

The AG-UI client works with WeChat Mini Programs by using the @cloudbase/agent-ui-miniprogram package to connect to the agent server and consume streamed AG-UI events for UI integration.

How do I route client-defined tools in an AG-UI TypeScript agent?

You route client-defined tools in an AG-UI TypeScript agent through adapter-specific setups like configuring clientTools routing, enabling the UI client to execute tools and return results back to the server.