# Google A2A — Agent-to-Agent Communication Protocol

> Open protocol by Google for AI agents to discover, authenticate, and communicate with each other. Enables multi-agent systems across different frameworks and providers. 10,000+ stars.

## Install

Save in your project root:

## Quick Use

```python
# Install
pip install a2a-sdk

# Create an A2A server (your agent becomes discoverable)
from a2a import A2AServer, AgentCard

card = AgentCard(
    name="code-reviewer",
    description="Reviews code for bugs and best practices",
    capabilities=["code_review", "security_audit"],
    endpoint="https://my-agent.example.com/a2a",
)

server = A2AServer(agent_card=card)

@server.on_task
async def handle_task(task):
    # Another agent sent you a task
    code = task.input["code"]
    review = await review_code(code)
    return {"review": review, "score": 8}

server.run(port=8080)
```

---

## Intro

Google A2A (Agent-to-Agent) is an open protocol for AI agents to discover, authenticate, and communicate with each other with 10,000+ GitHub stars. While MCP connects agents to tools, A2A connects agents to other agents — enabling multi-agent systems where a coding agent can delegate security review to a specialist agent, or a research agent can request data from an analytics agent. The protocol defines Agent Cards (discovery), task delegation, streaming responses, and authentication. Best for teams building multi-agent systems that need to interoperate across frameworks. Works with: any language, any agent framework. Setup time: under 5 minutes.

---

## How A2A Works

### The Problem
MCP connects agents to tools. But what about agent-to-agent communication?

```
Without A2A:
  Agent A (LangGraph) -- cannot talk to -- Agent B (CrewAI)

With A2A:
  Agent A (any framework) <-- A2A Protocol --> Agent B (any framework)
```

### Agent Cards (Discovery)
Every A2A agent publishes an Agent Card at `/.well-known/agent.json`:

```json
{
  "name": "security-auditor",
  "description": "Scans code for security vulnerabilities",
  "capabilities": ["security_audit", "dependency_scan", "owasp_check"],
  "endpoint": "https://security-agent.example.com/a2a",
  "authentication": { "type": "bearer" }
}
```

Other agents discover capabilities by fetching Agent Cards.

### Task Delegation

```python
from a2a import A2AClient

client = A2AClient("https://security-agent.example.com/a2a")

# Send a task to the security agent
result = await client.send_task({
    "type": "security_audit",
    "input": {"code": "...", "language": "python"}
})
# result: {"vulnerabilities": [...], "risk_score": "low"}
```

### Streaming Responses

```python
async for chunk in client.stream_task({"type": "code_review", "input": {...}}):
    print(chunk)  # Real-time updates from the remote agent
```

### Multi-Agent Architecture

```
Orchestrator Agent
  |-- A2A --> Code Writer Agent (writes code)
  |-- A2A --> Security Agent (reviews for vulnerabilities)
  |-- A2A --> Test Agent (generates and runs tests)
  |-- A2A --> Deploy Agent (deploys to production)
```

### A2A vs MCP

| Protocol | Connects | Direction |
|----------|----------|-----------|
| MCP | Agent <-> Tool | Agent calls tools |
| A2A | Agent <-> Agent | Agents collaborate |

They are complementary:
- MCP: Agent uses a GitHub tool to create a PR
- A2A: Agent asks another agent to review the PR

### Key Stats
- 10,000+ GitHub stars
- By Google DeepMind
- Agent discovery via Agent Cards
- Task delegation with streaming
- Framework-agnostic

### FAQ

**Q: What is A2A?**
A: An open protocol by Google for AI agents to discover and communicate with each other, enabling multi-agent systems across different frameworks.

**Q: Is A2A free?**
A: Yes, open protocol and SDK under Apache 2.0.

**Q: A2A vs MCP?**
A: MCP connects agents to tools. A2A connects agents to other agents. Use both together for full agent ecosystems.

---

## Source & Thanks

> Created by [Google](https://github.com/google). Licensed under Apache 2.0.
>
> [A2A](https://github.com/google/A2A) — stars 10,000+

Thanks to Google for standardizing how agents talk to each other.

---

<!-- ZH -->

## 快速使用

```python
pip install a2a-sdk
```

```python
from a2a import A2AClient
client = A2AClient("https://security-agent.example.com/a2a")
result = await client.send_task({"type": "security_audit", "input": {"code": "..."}})
```

---

## 简介

Google A2A（Agent-to-Agent）是一个开放协议，GitHub 10,000+ stars。让不同框架的 AI Agent 相互发现、认证和通信。MCP 连接 Agent 和工具，A2A 连接 Agent 和 Agent。适合构建需要跨框架互操作的多 Agent 系统。

---

## 来源与感谢

> Created by [Google](https://github.com/google). Licensed under Apache 2.0.
>
> [A2A](https://github.com/google/A2A) — stars 10,000+

---
Source: https://tokrepo.com/en/workflows/165627cf-cebe-47d3-9293-f75353499fd9
Author: AI Open Source