A2A
Agent-to-Agent (A2A) is a communication protocol for multi-agent system collaboration, designed to standardize task delegation, information exchange, and coordinated execution between AI agents. The protocol defines standard interfaces and message formats for inter-agent interaction, enabling heterogeneous agents to discover one another and collaborate across platforms and frameworks. It is typically used alongside tool-calling protocols such as MCP to form the foundational communication layer of multi-agent applications.
Core Facts
A2A (Agent-to-Agent) is an inter-agent communication protocol proposed by Google.
A2A introduces the concept of an Agent Card, where each Agent publishes a structured capability declaration document describing the task types it can handle, input/output formats, and authentication requirements.
The MCP protocol uses JSON-RPC 2.0 as its underlying communication protocol and defines three core primitives: Resources, Tools, and Prompts.
Timeline (last 90 days)
The underlying key protocol of SAP Agent Gateway is A2A (Agent-to-Agent).
In the A2A specification, each agent exposes its own information through a standardized agent card endpoint typically located at /.well-known/agent.json.
A2A emphasizes enterprise-grade authentication and authorization, supporting mechanisms such as OAuth 2.0.
In AI Agent Studio, ServiceNow allows third parties to access ServiceNow AI agents via the A2A protocol (version 0.3).
A2A version 0.3 indicates that the protocol is still iterating rapidly and carries evolution risk.
If an enterprise needs to connect with multiple third-party agent platforms, or is in a heavily regulated industry such as finance or healthcare, ServiceNow's A2A + AI Control Tower combination provides a more structured governance framework.
The engineering challenges of protocol convergence include identity mapping, permission propagation and audit trails, requiring identity conversion and permission narrowing at every hop.
MCP connects AI systems to capabilities, while A2A coordinates independent agents.
Conflating MCP and A2A commonly leads to two outcomes: a simple integration that should have been solved with tool calls is over-engineered into a multi-agent system, or a task that should be autonomously handled by independent agents is crammed into a single application that manages all tools.
Both MCP and A2A need authentication, authorization, timeout control and audit trails, and these engineering-level requirements do not disappear just because the protocols differ.
40 more timeline events
All Facts (20)
A2A (Agent-to-Agent) is an inter-agent communication protocol proposed by Google.
80%VerifiedA2A introduces the concept of an Agent Card, where each Agent publishes a structured capability declaration document describing the task types it can handle, input/output formats, and authentication requirements.
80%VerifiedThe MCP protocol uses JSON-RPC 2.0 as its underlying communication protocol and defines three core primitives: Resources, Tools, and Prompts.
80%VerifiedThe A2A protocol focuses on peer-to-peer collaboration between agents, while Anthropic's MCP focuses on connecting agents to tools and data sources.
80%VerifiedGoogle released the A2A (Agent-to-Agent) protocol in April 2025, specifically to address collaboration issues in multi-agent systems.
80%VerifiedThe A2A (Agent-to-Agent) protocol manages collaboration between Agents, addressing issues of capability discovery, task delegation, and collaborative interoperability.
75%VerifiedMCP addresses the connectivity problem (how a model reaches tools), while Skill addresses the usage problem (when and with what strategy the model combines tools).
75%VerifiedAn Agent Card is a standardized business card for agents in the A2A protocol, declaring capabilities, input/output schemas, authentication methods, and other information in JSON format, without exposing the internal model and tool implementations.
70%VerifiedThe A2A (Agent2Agent) protocol is a set of inter-agent communication specifications built on mature web standards, with the goal of enabling develop-once, connect-anywhere integration rather than being an entirely new underlying protocol.
65%VerifiedThe core design principle of the A2A protocol is that agents collaborate based on declared capabilities rather than inspecting each other's internal implementations.
65%VerifiedA2A and MCP are complementary rather than competitive, with MCP used at the lower layer to connect tools and data, and A2A used at the upper layer for multi-agent orchestration.
65%VerifiedThe core working unit of the A2A protocol is a Task, which has a unique ID and a complete state machine, including states such as submitted, working, completed, failed, canceled, and requires additional information.
65%VerifiedProtobuf is a language-neutral, platform-neutral structured data serialization protocol developed by Google, offering smaller data size and faster parsing speed compared to JSON and XML.
65%Unverified文章介绍的六大AI智能体协议为MCP、A2A、UCP、AP2、A2UI和AGUI
95%UnverifiedA2A协议通过Agent Card让每个Agent暴露能力清单，实现Agent之间零代码、零API契约的即插即用连接
90%UnverifiedGoogle创建了A2A（Agent-to-Agent）协议并将其捐赠给Linux基金会
90%UnverifiedA2A协议的传输层使用JSON-RPC 2.0走HTTPS，长任务用Server-Sent Events做流式更新或Webhook做异步通知，版本0.3还加了gRPC支持
90%UnverifiedA2A协议基于HTTP/SSE构建，引入了Agent Card概念，每个Agent发布描述自身能力的元数据卡片
90%UnverifiedA2A和MCP在2025年底被捐赠给Linux Foundation旗下的Agentic AI Foundation共同治理
80%UnverifiedA2A协议在2026年4月正式推至1.0版本
80%