A2A Samples for Azure AI Foundry Agent SDK

This directory contains three comprehensive examples demonstrating how to integrate Azure AI Foundry Agent Service with Google's Agent-to-Agent (A2A) Protocol. These samples showcase different approaches to building intelligent agents using Azure's AI services, from simple calendar management to sophisticated multi-agent orchestration systems.

🔋 Core Technologies

• Azure AI Foundry Agent Service: Intelligent agent capabilities with Azure AI
• Google A2A SDK: Agent-to-agent communication framework
• Model Context Protocol (MCP): Standardized tool communication
• Azure Functions: Serverless hosting for MCP services

📁 Examples Overview

1. Azure Foundry Agent (./azurefoundryagent)

A calendar management agent that demonstrates core Azure AI Foundry integration with A2A protocol.

Key Features:

• 🤖 AI Foundry Integration: Build intelligent agents using Azure AI Foundry
• 📅 Calendar Management: Check schedule availability, get upcoming events
• 🔄 A2A Framework: Support agent-to-agent communication and collaboration
• 💬 Conversation Capabilities: Natural language processing and multi-turn conversations
• 🛠️ Tool Integration: Simulated calendar API tool integration

Use Cases:

• "Check my calendar for tomorrow"
• "What meetings do I have this week?"
• "Am I available on Friday afternoon?"

Technologies:

• Azure AI Foundry Agent Service
• Azure AI Projects SDK
• A2A SDK for Python
• Starlette web framework

2. Currency Agent Demo (./currencyagentdemo)

A comprehensive currency exchange system combining Azure AI Foundry, MCP services, and A2A protocol for real-time currency conversion.

Architecture Components:

1. 🔌 MCP Server: Azure Functions-based service providing currency exchange tools
2. 💱 Currency Agent: Azure AI Foundry agent integrated with A2A protocol

Key Features:

• 🎯 Azure AI Agent Service: Leverages Azure AI Foundry for intelligent responses
• 🔧 Model Context Protocol (MCP): Standardized tool communication protocols
• 🤝 Google A2A SDK: Agent-to-agent communication framework
• ☁️ Azure Functions: Serverless MCP service hosting
• 💰 Real-time Exchange Rates: Uses Frankfurter API for live currency data
• 📡 Streaming Responses: Real-time response streaming

Available Tools:

• hello_mcp: Connectivity test tool
• get_exchange_rate: Currency conversion with currency_from and currency_to parameters

Use Cases:

• "Convert 100 USD to EUR"
• "What's the current exchange rate from GBP to JPY?"
• "How much is 50 CAD in Australian dollars?"

Technologies:

• Azure AI Foundry Agent Service
• Azure Functions (for MCP server)
• Model Context Protocol (MCP)
• A2A SDK for Python
• Frankfurter API for exchange rates

3. Multi-Agent System (./multi_agent)

A sophisticated multi-agent architecture that demonstrates intelligent task routing and delegation to specialized remote agents using Azure AI Foundry, A2A protocol, and Semantic Kernel.

Architecture Components:

1. 🎯 Host Agent: Central routing system powered by Azure AI Foundry
2. 🤖 Remote Agents: Specialized task executors (Playwright, Tool agents)
3. 🔌 MCP Server: Azure Functions-based service providing extensible functionality
4. 🧠 Semantic Kernel: Advanced agent framework for intelligent routing

Key Features:

• 🎯 Intelligent Routing: Azure AI Foundry-powered central agent for task delegation
• 🤝 Multi-Agent Coordination: Agent-to-agent communication using A2A protocol
• 🧠 Semantic Kernel Integration: Advanced semantic understanding and routing
• 🌐 Web Interface: Modern Gradio-based chat interface with real-time streaming
• 🔧 Playwright Integration: Web automation and browser-based task execution
• ☁️ MCP Azure Functions: Serverless tool integration with Model Context Protocol
• 📡 Multiple Communication Protocols: STDIO, SSE, and A2A protocol support

Available Agent Types:

• Playwright Agent: Web automation and browser tasks
• Tool Agent: General-purpose tool execution

Technologies:

• Azure AI Foundry Agent Service
• Semantic Kernel
• A2A SDK for Python
• Model Context Protocol (MCP)
• Azure Functions
• Playwright for web automation
• Gradio for web interface

🚀 Getting Started

Prerequisites

• Python 3.12+ (azurefoundryagent) or Python 3.13+ (currencyagentdemo)
• Azure AI Foundry project and deployment
• Azure subscription (for Functions in currency demo)
• UV package manager (recommended)

Quick Setup

1. Choose an example directory
2. Copy .env.template to .env and configure Azure settings
3. Install dependencies: uv sync
4. Run the agent: uv run .

Required Environment Variables

AZURE_AI_FOUNDRY_PROJECT_ENDPOINT=Your Azure AI Foundry Project Endpoint
AZURE_AI_AGENT_MODEL_DEPLOYMENT_NAME=Your Azure AI Foundry Deployment Model Name

🎯 When to Use Each Example

Use Azure Foundry Agent when you want to:

• Learn core Azure AI Foundry + A2A integration
• Build simple tool-based agents
• Understand basic calendar/scheduling functionality
• Get started with minimal setup

Use Currency Agent Demo when you want to:

• Build production-ready agents with external services
• Implement MCP protocol with Azure Functions
• Create agents that interact with real-time APIs
• Understand complex multi-service architecture