프로젝트 문서

Spring AI Alibaba

A production-ready framework for building Agentic, Workflow, and Multi-agent applications.

Agent Framework Docs, Graph Docs, Spring AI, Examples.

Architecture

Spring AI Alibaba Admin is a one-stop Agent platform that supports visualized Agent development, observability, evaluation, and MCP management, etc. It also integrates with open-source low-code platforms like Dify, enabling rapid migration from DSL to Spring AI Alibaba project.

Spring AI Alibaba Agent Framework is an agent development framework that can quickly develop agents with built-in Context Engineering and Human In The Loop support. For scenarios requiring more complex process control, Agent Framework offers built-in workflows like SequentialAgent, ParallelAgent, RoutingAgent, LoopAgent.

Spring AI Alibaba Graph serves as the underlying runtime of the Agent Framework, providing essential capabilities such as persistence, workflow orchestration, and streaming required for long-running stateful agents. Compared to the Agent Framework, users can build more flexible multi-agent workflows based on the Graph API.

Core Features

Multi-Agent Orchestration: Compose multiple agents with built-in patterns including SequentialAgent, ParallelAgent, RoutingAgent, and LoopAgent for complex task execution.
Multimodal Support: ReactAgent with text and media input (image understanding). ReactAgent with tool based image or audio generation.
Voice Agent: WebSocket-based real-time voice agent that supports streaming audio or text input and responds with generated audio.
Context Engineering: Built-in best practices for context engineering policies to improve agent reliability and performance, including human-in-the-loop, context compaction, context editing, model & tool call limit, tool retry, planning, dynamic tool selection.
Graph-based Workflow: Graph based workflow runtime and api for conditional routing, nested graphs, parallel execution, and state management. Export workflows to PlantUML and Mermaid formats.
A2A Support: Agent-to-Agent communication support with Nacos integration, enabling distributed agent coordination and collaboration across services.
Rich Model, Tool and MCP Support: Leveraging core concepts of Spring AI, supports multiple LLM providers (DashScope, OpenAI, etc.), tool calling, and Model Context Protocol (MCP).
One-stop Agent Platform: Build agent in a visualized way, deploy agent without code or export as a standalone java project.

Getting Started

Prerequisites

Requires JDK 17+.
Choose your LLM provider and get the API-KEY.

Quickly Run a ChatBot

There's a ChatBot example provided by the community at examples/chatbot.

Download the code.

git clone --depth=1 https://github.com/alibaba/spring-ai-alibaba.git
cd spring-ai-alibaba

Start the ChatBot.

Before starting, set API-KEY first (visit Aliyun Bailian to get API-KEY):

# this example uses 'spring-ai-alibaba-starter-dashscope', visit https://java2ai.com to learn how to use OpenAI/DeepSeek.
export AI_DASHSCOPE_API_KEY=your-api-key

# Maven installation is optional when using mvnw.
./mvnw -pl examples/chatbot spring-boot:run

Chat with ChatBot.

Open the browser and visit http://localhost:8080/chatui/index.html to chat with the ChatBot.

Chatbot Code Explained

Add dependencies

<dependencies>
  <dependency>
    <groupId>com.alibaba.cloud.ai</groupId>
    <artifactId>spring-ai-alibaba-agent-framework</artifactId>
    <version>1.1.2.0</version>
  </dependency>
  <!-- Assume you are going to use DashScope Model. Refer to docs for how to choose model.-->
  <dependency>
    <groupId>com.alibaba.cloud.ai</groupId>
    <artifactId>spring-ai-alibaba-starter-dashscope</artifactId>
    <version>1.1.2.1</version>
  </dependency>
</dependencies>

Define Chatbot

For more details of how to write a Chatbot, please check the Quick Start on our official website.

📚 Documentation

Overview - High level overview of the framework
Quick Start - Get started with a simple agent
Agent Framework Tutorials - Step by step tutorials
Use Graph API to Build Complex Workflows - In-depth user guide for building multi-agent and workflows
Spring AI Basics - Ai Application basic concepts, including ChatModel, MCP, Tool, Messages, etc.
Chat Memory - Spring AI reference for chat memory repositories and usage

Project Structure

This project consists of several core components:

spring-ai-alibaba-agent-framework: A multi-agent framework designed for building intelligent agents with built-in context engineering best practices.
spring-ai-alibaba-graph: The underlying runtime for Agent Framework. We recommend developers to use Agent Framework but it's totally fine to use the Graph API directly.
spring-ai-alibaba-admin: A one-stop Agent platform that supports visualized Agent development, observability, evaluation, and MCP management, etc.
spring-ai-alibaba-studio: The embedded ui for quickly debugging agent in a visualized way.
spring-boot-starters: Starters integrating Agent Framework with Nacos to provide A2A and dynamic config features.