Table of Content
AI Information
AI Tutorial
AI Agent
Latest Information
Hot Information
LLMs Information
Technical Frontiers
Industry Applications
Prompt Tutorial
AI Programming
AI Video Generation
AI Audio Generation
Knowledge Base Setup
AI Development Guides
Industry Tutorials
Practical Cases
AI Resource Library
Agent Information
Agent Setup
Agent Training

Google releases A2A protocol: opens a new era of AI agent collaboration and fully supports MCP

Written by
Caleb Hayes
Updated on:July-02nd-2025
Recommendation

Google and DeepMind work together to lead new breakthroughs in the field of AI. The A2A protocol facilitates multi-agent collaboration.

Core content:
1. The A2A protocol complements the MCP and jointly promotes cross-platform intelligent agent collaboration
2. The five design principles of the A2A protocol and its working mechanism
3. Example demonstration: The application of the A2A protocol in candidate screening and future roadmap

Yang Fangxian
Founder of 53AI/Most Valuable Expert of Tencent Cloud (TVP)

Google released  the Agent2Agent Protocol (A2A) this morning. A2A is a key step towards  universal intelligent agent collaboration  . It enables cross-platform and cross-vendor AI agents to truly work together, marking a leap from "single-agent intelligence" to "multi-agent system intelligence."

Link? -  https://google.github.io/A2A/

The Agent-to-Agent (A2A) protocol  is a standard jointly proposed by Google and DeepMind to build a communication and collaboration system between remote intelligent agents.

MCP (Multi-Agent Collaboration Protocol)  is fully supported, and  A2A is its supplementary protocol to address the shortcomings of MCP in practical applications.

? A2A and MCP (Anthropic) relationship

  • It is not competition, but  complementarity . The goal is to solve the problem of lack of standards in agent calls and promote cross-platform compatibility.
  • A2A focuses on  Agent-to-Agent collaboration
  • MCP provides  context and tool standards (Model Context Protocol)
  • The combination of the two can build a truly "end-to-end autonomous" agent ecosystem.

? Five design principles of A2A protocol

in principle
illustrate
1. Embrace agent intelligence
Supports multi-agent collaboration without shared memory, context or tools
2. Based on existing standards
Built on widely used standards such as HTTP, SSE, JSON-RPC, etc.
3. Secure by default
Enable an enterprise certification system at the same level as OpenAPI
4. Support long tasks and state management
Support for asynchronous, persistent tasks (e.g. candidate screening, background checks)
5. Modality is irrelevant
Support multi-modal collaboration such as text, audio, and video

How A2A Works

A2A defines two types of roles:

  • Client Agent : Initiate tasks and assign collaboration
  • Remote Agent : Receives tasks and returns results (called Artifact)

Key components:

  • Agent Card : A capability description file in JSON format (the "business card" of the agent)
  • Task Object : Description of the task life cycle, including status updates, completion markers, etc.
  • Message communication : used to transmit context, results, user instructions, etc.
  • User Experience Negotiation : Negotiate content formats for end-user interface capabilities (video/iframe, etc.)

Example demonstration: Candidate screening

On a platform like  Agentspace  :

  1. The user (HR) entrusts the main agent to find candidates for a certain position
  2. The master agent calls multiple remote talent matching agents to collect resumes
  3. After the user confirms, interviews and background checks are automatically arranged
  4. The entire process does not require manual cross-system switching, greatly improving recruitment efficiency

Related PostsView More
《Can’t tell the difference between RAG, MCP, Agent and AI concepts? This article will turn newbies into experts in seconds! 》
Explore the mysterious technology in the field of AI, and learn about RAG in one article! Today, we will explain the RAG technology and its principles in detail. RAG is like a "super library" for AI assistants, providing accurate basis for answering questions. Say goodbye to the confusion about RAG and easily master this cutting-edge technology. Click to read for details!
Clara Bennett
Survival Guide in the AI ​​Era: Why Your Experience Is More Valuable Than Knowledge?
Explore the development secrets of the AI ​​era and deeply analyze the status quo and principles of big model technology. The article focuses on the computing power, data, reasoning and other challenges faced by big models, as well as the current development direction, such as optimizing the underlying computing power, innovating architecture algorithms, and embedding business closed loops. It...
Jasper Cole
Cline v3.15 Updates and Highlights
Cline v3.15 is a major update, bringing 18 new features, 10 updates, and 5 fixes. It optimizes file editing interactions, enhances task transparency, reduces usage costs, and promotes community collaboration through open source documentation. This version reflects the continuous progress of AI big models in application development, involving innovations in model fine-tuning methods and...
Silas Grey