Key Points
- A2A support in Spring AI Alibaba: The framework adds A2A (Agent-to-Agent) protocol client support, enabling remote, cross-process and cross-service agent interactions beyond in-process multi-agent setups.
- Implementation: Delivered through the A2A protocol client in
spring-ai-alibaba-graph-core, leveraging dependencies from the A2A SDK (a2a-java-sdk-client,a2a-java-sdk-server-common). Agents discover and invoke remote agents via AgentCards. - Timeline: A2A collaboration was planned after the 1.0 GA release (June 2025); a client adaptation was merged via pull request #2254 in August 2025, with
spring-ai-alibaba-a2a-commonavailable up to version 1.0.0.4 as of September 2025. - Benefits: Enables enterprise-level distributed AI applications such as remote multi-agent workflows with automatic request handling and state synchronization, improving scalability and interoperability.
- GitHub: https://github.com/alibaba/spring-ai-alibaba
- Documentation: https://java2ai.com/en/docs/
- Spring AI Alibaba 1.0 GA release
- Google's Agent2Agent (A2A) protocol announcement coverage
- PR #2254: Add A2A Protocol Client Adaptation
- spring-ai-alibaba-a2a module
- Spring AI Alibaba official site
Overview of Spring AI Alibaba
Spring AI Alibaba is an open-source agentic AI framework built on Spring AI, tailored for Java developers building ChatBots, workflows, and multi-agent systems. It integrates with Alibaba Cloud's Bailian platform (LLM access, RAG) and enterprise tools like Nacos (service discovery) and ARMS (observability). Graph-based orchestration via Spring AI Alibaba Graph allows declarative definitions of nodes, edges, and states.
Resources:
The 1.0 GA release (June 2025) introduced production-ready graph-based multi-agent frameworks, streaming support, human-in-the-loop interventions, and persistent memory, with pre-built nodes such as LlmNode and ToolNode, plus workflow export to PlantUML or Mermaid.
Understanding the A2A Protocol
A2A is an open standard proposed by Google in 2024 for autonomous agent-to-agent interaction. Agents discover capabilities via Agent Cards (JSON manifests describing capabilities, inputs, and outputs), delegate tasks, and communicate statelessly over HTTP, JSON-RPC, and OpenAPI-style authentication. It supports multiple modalities (text, audio, video) and long-running tasks with real-time feedback. In Spring AI Alibaba, A2A complements MCP (Model Context Protocol), which handles secure context exchange with external systems — a potential "MCP+A2A" standard combination.
Integration and Usage
The A2A client adaptation enables remote agent calling across distributed services. Developers use classes such as A2aRemoteAgent and A2aNode, which handle URL-based AgentCard parsing, request formatting, and response integration into graph state. A remote writer agent can be configured with a builder pattern (A2aRemoteAgent.builder()), maintaining API consistency with local agents. In distributed setups, Nacos serves as the registry for dynamic agent discovery and routing.
Release Progression
| Version/Component | Release Date | Key A2A Features | Related Modules | |-------------------|--------------|------------------|-----------------| | 1.0 GA (Core Framework) | June 2025 | Planned A2A participation; no direct implementation yet | spring-ai-alibaba-graph-core | | Pull Request #2254 | August 2025 (merged) | 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 | spring-ai-alibaba-a2a-common (Apache 2.0) | | Starter A2A Server | September 2025 | Server-side A2A support | spring-ai-alibaba-starter-a2a-server |
Future Developments
Current work focuses on client-side adaptations, with ongoing community efforts co-building protocols like MCP and AG-UI toward zero-code agent development platforms. Check the Maven repository for the latest releases. Related community products include JManus (plan-act agents) and Spring AI Alibaba Studio for evaluation and deployment.