A2A Protocol Support Documentation

This document explains how to integrate and use the A2A (Agent-to-Agent) protocol in the Agentscope Runtime SDK, allowing your agent to support the A2A standard and automatically generate the agent card.

1. Background & Purpose

A2A (Agent-to-Agent) protocol defines a standard interface for interaction between intelligent agents (including action, card, metadata, etc.). By wrapping your agent with A2AFastAPIDefaultAdapter, you can instantly enable A2A protocol support and automatically generate the agent card.

2. Main Classes and Methods

• A2AFastAPIDefaultAdapter(agent): Wraps an existing agent (e.g., LLMAgent) as an A2A protocol-compliant server.

• protocol_adapters: A list specifying supported protocol adapters when deploying the runner, including the A2A protocol adapter.

3. Integration Steps

Below are the key steps you need to add A2A protocol support to your agent:

Step 1: Create your agent instance

llm_agent = LLMAgent(
    model=QwenLLM(),
    name="llm_agent",
    description="A simple LLM agent to generate a short story",
)

Step 2: Wrap agent with A2A protocol adapter

a2a_protocol = A2AFastAPIDefaultAdapter(agent=llm_agent)

Step 3: Build context management services

session_history_service = InMemorySessionHistoryService()
memory_service = InMemoryMemoryService()
context_manager = ContextManager(
    session_history_service=session_history_service,
    memory_service=memory_service,
)

Step 4: Build runner

runner = Runner(
    agent=llm_agent,
    context_manager=context_manager,
)

Step 5: Deploy runner with the protocol_adapters argument

deploy_manager = LocalDeployManager(host="localhost", port=server_port)

deployment_info = await runner.deploy(
    deploy_manager,
    endpoint_path=f"/{server_endpoint}",
    protocol_adapters=[a2a_protocol],  # PROTOCOL ADAPTERS declaration
)

print("✅ Service deployed successfully!")
print(f"URL: {deployment_info['url']}/{server_endpoint}")

4. Complete Example

The following example demonstrates how to deploy a local agent service that supports A2A protocol:

import asyncio
import os

from agentscope_runtime.engine import Runner, LocalDeployManager
from agentscope_runtime.engine.agents.llm_agent import LLMAgent
from agentscope_runtime.engine.deployers.adapter.a2a import A2AFastAPIDefaultAdapter
from agentscope_runtime.engine.llms import QwenLLM
from agentscope_runtime.engine.services.context_manager import ContextManager
from agentscope_runtime.engine.services.memory_service import InMemoryMemoryService
from agentscope_runtime.engine.services.session_history_service import InMemorySessionHistoryService

async def main():
    # Read environment variables
    server_port = int(os.environ.get("SERVER_PORT", "8090"))
    server_endpoint = os.environ.get("SERVER_ENDPOINT", "agent")

    # Step 1: Create agent instance
    llm_agent = LLMAgent(
        model=QwenLLM(),
        name="llm_agent",
        description="A simple LLM agent to generate a short story",
    )

    # Step 2: Wrap agent with A2A protocol adapter
    a2a_protocol = A2AFastAPIDefaultAdapter(agent=llm_agent)

    # Step 3: Build context management services
    session_history_service = InMemorySessionHistoryService()
    memory_service = InMemoryMemoryService()
    context_manager = ContextManager(
        session_history_service=session_history_service,
        memory_service=memory_service,
    )

    # Step 4: Build runner and deploy
    runner = Runner(
        agent=llm_agent,
        context_manager=context_manager,
    )

    deploy_manager = LocalDeployManager(host="localhost", port=server_port)

    deployment_info = await runner.deploy(
        deploy_manager,
        endpoint_path=f"/{server_endpoint}",
        protocol_adapters=[a2a_protocol],  # PROTOCOL ADAPTERS declaration
    )

    print("✅ Service deployed successfully!")
    print(f"URL: {deployment_info['url']}/{server_endpoint}")

asyncio.run(main())

5. Notes

Wrapping your agent with A2AFastAPIDefaultAdapter and passing it through protocol_adapters are both required for enabling the A2A protocol and auto-generating your agent card endpoint.