spring-ai-alibaba

Spring AI Alibaba is Alibaba's production-ready framework for building agentic, workflow and multi-agent applications in Java. It builds on Spring AI concepts (ChatModel, tools, MCP, messages) and consists of three layers: an Agent Framework with built-in context engineering and human-in-the-loop support, a Graph runtime underneath that provides persistence, workflow orchestration and streaming for long-running stateful agents, and Spring AI Alibaba Admin, a platform for visual agent development, observability, evaluation and MCP management.

It suits Java and Spring Boot teams who want agents inside their existing services rather than in a separate Python stack. Multi-agent orchestration comes with built-in patterns such as sequential and parallel agents; ReAct agents support text and image input and tool-based image or audio generation; a WebSocket-based realtime voice agent is included; and A2A support with Nacos integration enables distributed agent collaboration. It supports multiple model providers, including Alibaba Cloud DashScope, OpenAI and DeepSeek.

Getting started: requires JDK 17+ and an API key from your model provider. Clone the repository and run the chatbot example with `./mvnw -pl examples/chatbot spring-boot:run`, or add the `spring-ai-alibaba-agent-framework` dependency and a model starter to your Spring Boot project.