Python A2A: Agent-to-Agent Protocol Library

Building interoperable AI agent ecosystems through protocol standardization

Python A2A is a comprehensive implementation of Google's Agent-to-Agent (A2A) protocol with Model Context Protocol (MCP) integration, enabling developers to build sophisticated multi-agent systems that can collaborate to solve complex problems while maintaining compatibility with the broader AI agent ecosystem.

Key Features

• 🌐 Complete A2A Protocol Implementation: Full compliance with Google's A2A specification
• 🔌 First-class MCP Support: Native integration with Model Context Protocol
• 🔄 Bidirectional LangChain Integration: Seamless interoperability with LangChain ecosystem
• 🌊 Advanced Streaming: Real-time response streaming with error recovery
• 🔍 Agent Discovery: Built-in registry and discovery mechanisms
• 🔄 Workflow Engine: Orchestrate complex agent interactions with conditional logic
• 🚀 Framework Agnostic: Works with Flask, FastAPI, Django, or standalone
• 🧩 LLM Provider Flexibility: Native support for OpenAI, Anthropic, AWS Bedrock
• ⚠️ Production-Ready Error Handling: Comprehensive error handling for robustness

Why Python A2A?

• Standardization: Implement the official A2A specification for future-proof agent systems
• Interoperability: Connect seamlessly with different agent ecosystems
• Flexibility: Support for various LLM providers and Python frameworks
• Modularity: Compose agents with specialized capabilities into complex systems
• Developer Experience: Comprehensive documentation and working examples
• Enterprise Ready: Production-quality implementation with robust error handling

Getting Started

To install the library:

pip install python-a2a  # Basic installation
pip install "python-a2a[all]"  # All features

For more installation options, see the Installation Guide.

Quick Example: Creating an A2A Agent

from python_a2a import A2AServer, skill, agent, run_server

@agent(
    name="Weather Agent",
    description="Provides weather information"
)
class WeatherAgent(A2AServer):
    
    @skill(
        name="Get Weather",
        description="Get current weather for a location"
    )
    def get_weather(self, location):
        return f"It's sunny and 75°F in {location}"
    
    def handle_task(self, task):
        return {"output": self.get_weather(task.input)}

# Run the server
agent = WeatherAgent()
run_server(agent, port=5000)

Quick Example: Connecting to an Agent

from python_a2a import HTTPClient

client = HTTPClient("http://localhost:5000")
response = client.send_message("What's the weather in New York?")
print(response.content)

For more examples, see the Examples Guide.

Documentation Structure

• Installation - Complete installation instructions
• Quickstart - Get up and running quickly
• Core Concepts - A2A protocol fundamentals
• Examples - Practical, working examples
• Agent Creation - Building custom A2A agents
• Client Usage - Connecting to and using agents
• Streaming - Real-time response streaming
• MCP Integration - Using the Model Context Protocol
• LangChain Integration - Interoperability with LangChain
• Agent Networks - Building ecosystems of agents
• Advanced Usage - Workflows, discovery, and more
• API Reference - Detailed API documentation
• Contributing - Guide for contributors

Use Cases

• Research Assistants: Network of specialized knowledge agents
• Enterprise Systems: AI orchestration across departments
• Customer Service: Multi-step assistants with specialized capabilities
• Education: Agent collaboration for learning environments
• Development: Rapid prototyping of AI-powered systems

License

This project is licensed under the MIT License - see the LICENSE file for details.