### Key Points - **Support Status**: Spring AI Alibaba provides support for the A2A (Agent-to-Agent) protocol, enabling remote distributed agent architectures for multi-agent collaboration across services. This integration facilitates cross-process and cross-service agent interactions, moving beyond in-process limitations. - **Implementation Details**: The support is implemented through the A2A protocol client in the spring-ai-alibaba-graph-core module, allowing agents to discover and invoke remote agents via AgentCards. It leverages dependencies from the A2A SDK for seamless communication. - **Availability**: Introduced after the 1.0 GA release in June 2025, with components like spring-ai-alibaba-a2a-common available in versions up to 1.0.0.4 as of September 2025. - **Purpose and Benefits**: This enables enterprise-level distributed AI applications, such as remote multi-agent workflows, with features like automatic request handling and state synchronization, enhancing scalability and interoperability. ### Overview of Spring AI Alibaba Spring AI Alibaba is an open-source agentic AI framework built on Spring AI, tailored for Java developers to create ChatBots, workflows, and multi-agent systems. It integrates deeply with Alibaba Cloud's Bailian platform for large language models (LLMs), Retrieval-Augmented Generation (RAG), and other AI services, while supporting enterprise tools like Nacos for service discovery and ARMS for observability. The framework emphasizes graph-based orchestration via Spring AI Alibaba Graph, allowing declarative definitions of nodes, edges, and states for complex AI flows. For more details, visit the official GitHub repository at https://github.com/alibaba/spring-ai-alibaba or the documentation at https://java2ai.com/en/docs/. ### Understanding A2A Protocol A2A, or Agent-to-Agent protocol, is an open standard proposed by Google in 2024 for enabling autonomous interactions between AI agents. It allows agents to discover capabilities via "Agent Cards" (JSON descriptions), delegate tasks, and communicate in a stateless, secure manner using standards like HTTP, JSON-RPC, and OpenAPI authentication. Designed for multi-modal support (text, audio, video) and long-running tasks, A2A promotes interoperability across platforms, breaking down silos in enterprise environments. In Spring AI Alibaba, A2A complements protocols like MCP (Model Context Protocol) for context sharing with external systems, potentially becoming a standard combination for AI ecosystems. ### Integration and Usage In Spring AI Alibaba, A2A support is added to enable remote agent calling, addressing scenarios where agents operate across distributed services. Developers can build remote agents using classes like A2aRemoteAgent and A2aNode, which handle URL-based AgentCard parsing, request formatting, and response integration into the overall state. For example, a remote writer agent can be configured with a simple builder pattern, maintaining API consistency with local agents. This is particularly useful in distributed setups with Nacos for registry and routing. Community contributions, such as the merged pull request in August 2025, ensure ongoing enhancements. ### Future Developments While A2A was planned for deeper involvement post-1.0 GA, current implementations focus on client-side adaptations. Ongoing community efforts include co-building with protocols like MCP and AG-UI, aiming for zero-code agent development platforms. Users are encouraged to check the latest releases on Maven (e.g., https://mvnrepository.com/artifact/com.alibaba.cloud.ai/spring-ai-alibaba-a2a-common) for updates. --- Spring AI Alibaba represents a significant advancement in Java-based AI development, extending the core Spring AI project with Alibaba-specific integrations to support sophisticated agentic applications. At its foundation, the framework provides abstractions for building AI-enhanced systems, including ChatClients for single-agent interactions and StateGraphs for orchestrating multi-step workflows. These capabilities are bolstered by integrations with Alibaba Cloud services, such as the Bailian platform for accessing Tongyi Qianwen models, RAG for knowledge retrieval, and NL2SQL for natural language to database queries. The 1.0 GA release in June 2025 marked a milestone, introducing production-ready features like graph-based multi-agent frameworks, streaming support, human-in-the-loop interventions, and persistent memory storage. Developers can leverage pre-built nodes (e.g., LlmNode for model calls, ToolNode for function execution) and export workflows to visual formats like PlantUML or Mermaid for better debugging and collaboration. A key evolution in Spring AI Alibaba is its embrace of distributed architectures, where the A2A protocol plays a pivotal role. Originally proposed by Google DeepMind, A2A facilitates standardized communication between agents, allowing them to operate autonomously across different systems without tight coupling. Its design principles emphasize security, modality-agnostic interactions (supporting text, audio, and video), and handling of prolonged tasks with real-time feedback. In practice, A2A uses Agent Cards—JSON manifests that describe an agent's capabilities, inputs, and outputs—enabling dynamic discovery and task delegation. This protocol synergizes with Alibaba's MCP (Model Context Protocol), which focuses on secure context exchange between models and external APIs, potentially forming a "MCP+A2A" standard for comprehensive AI interoperability. The integration of A2A into Spring AI Alibaba was achieved through targeted enhancements, notably via a pull request merged in August 2025 that added A2A client adaptations to the graph-core module. This update shifts from purely in-process multi-agent setups to remote scenarios, where agents can invoke each other across services using URLs to fetch Agent Cards, parse them, and execute tasks with automatic format conversions for requests and responses. For instance, in a distributed setup, Nacos serves as the registry for dynamic agent discovery, combined with A2A for communication, ensuring load balancing and fault tolerance in enterprise deployments. Practical examples include building a remote writer agent that generates content based on inputs from other agents, all while maintaining consistency with Spring's bean lifecycle management. To illustrate the progression of A2A support, consider the following table summarizing key releases and features: | Version/Component | Release Date | Key A2A Features | Related Dependencies/Modules | |-------------------|--------------|------------------|------------------------------| | 1.0 GA (Core Framework) | June 2025 | Planned participation in A2A protocol co-building; no direct implementation yet | spring-ai-alibaba-graph-core (initial multi-agent support) | | Pull Request #2254 (Graph-Core Update) | August 2025 (Merged) | Added A2A client for remote agent interactions; AgentCard parsing, remote invocation | a2a-java-sdk-client, a2a-java-sdk-server-common | | 1.0.0.4 (A2A Common) | September 2025 | Stable A2A common utilities for distributed agents | spring-ai-alibaba-a2a-common (Apache 2.0 licensed) | | Starter A2A Server (Related Artifact) | September 2025 | Server-side support for A2A-enabled agents | spring-ai-alibaba-starter-a2a-server | This table highlights how A2A evolved from conceptual plans to concrete implementations, with dependencies ensuring compatibility. Developers starting with A2A in Spring AI Alibaba should include relevant starters in their Maven pom.xml, configure API keys, and define remote agents using builders like A2aRemoteAgent.builder(). For advanced use cases, such as multi-agent collaboration in research agents like DeepResearch, A2A enables tools like web crawling and Python scripting to be distributed across nodes. Looking ahead, the framework's roadmap includes deeper protocol integrations, visual debugging tools, and platforms like Spring AI Alibaba Studio for evaluation and deployment. Community-driven products, such as JManus for plan-act agents and AgentScope for scoping, demonstrate real-world applications. Overall, Spring AI Alibaba's A2A support positions it as a robust choice for building scalable, interoperable AI systems in Java, with strong ties to Alibaba's ecosystem for enterprise reliability. ### Key Citations - [Spring AI Alibaba 1.0 GA 正式发布，开启Java 智能体开发新时代！](https://java2ai.com/blog/spring-ai-alibaba-10-ga-release/) - [Spring AI Alibaba Graph 智能体开发指南与实践](https://juejin.cn/post/7542718732163399731) - [谷歌发布Agent2Agent (A2A) 协议，未来"MCP+A2A" 成标配？](https://java2ai.com/blog/google-a2a-protocol/) - [Spring AI Alibaba 发布企业级MCP 分布式部署方案](https://www.cnblogs.com/alisystemsoftware/p/18900511) - [阿里云开源AI 应用开发框架：Spring AI Alibaba](https://www.cnblogs.com/alisystemsoftware/p/18506747) - [spring-ai-alibaba-a2a](https://github.com/alibaba/spring-ai-alibaba/tree/main/spring-ai-alibaba-a2a) - [【Feature】Add A2A Protocol Client Adaptation to spring-ai-alibaba ...](https://github.com/alibaba/spring-ai-alibaba/pull/2254) - [spring-ai-alibaba-a2a-common](https://mvnrepository.com/artifact/com.alibaba.cloud.ai/spring-ai-alibaba-a2a-common) - [使用Spring AI Alibaba调用百炼大模型应用](https://help.aliyun.com/zh/model-studio/spring-ai-alibaba-integrate-llm-application) - [alibaba/spring-ai-alibaba: Agentic AI Framework for Java Developers](https://github.com/alibaba/spring-ai-alibaba) - [Spring AI Alibaba](https://sca.aliyun.com/en/ai/) - [Spring AI Alibaba 1.0 GA officially released, marking the advent of a ...](https://www.alibabacloud.com/blog/spring-ai-alibaba-1-0-ga-officially-released-marking-the-advent-of-a-new-era-in-java-agent-development_602299) - [Accelerating AI Integration with Alibaba Cloud's Model Context ...](https://www.alibabacloud.com/blog/accelerating-ai-integration-with-alibaba-cloud%25E2%2580%2599s-model-context-protocol-mcp_602490) - [Spring AI Alibaba Graph Workflow Quick Start](https://java2ai.com/en/docs/1.0.0.2/get-started/workflow/) - [Alibaba Cloud Open Source AI Application Development Framework ...](https://medium.com/%40lararoself/spring-ai-alibaba-alibaba-cloud-open-source-ai-application-development-framework-d8e0a56e02a9) - [spring-ai-alibaba/spring-ai-alibaba-admin](https://github.com/spring-ai-alibaba/spring-ai-alibaba-admin) - [Spring AI Alibaba: The Ultimate Java Framework for Building](https://www.linkedin.com/pulse/spring-ai-alibaba-java-developers-new-game-changer-apps-pasindu-1930c) - [Agentic AI with the Agent2Agent Protocol (A2A) and MCP using ...](https://www.kai-waehner.de/blog/2025/05/26/agentic-ai-with-the-agent2agent-protocol-a2a-and-mcp-using-apache-kafka-as-event-broker/) - [【Feature】Add A2A Protocol Client Adaptation to spring-ai-alibaba ...](https://github.com/alibaba/spring-ai-alibaba/pull/2254) - [What is the Agent-to-Agent Protocol (A2A)](https://codefarm0.medium.com/what-is-the-agent-to-agent-protocol-a2a-c80dabc6d82b) - [AI's big interoperability moment: Why A2A and MCP are key for ...](https://venturebeat.com/ai/ais-big-interoperability-moment-why-a2a-and-mcp-are-key-for-agent-collaboration)