examples/

A2A ADK Examples

This directory contains scenario-based examples demonstrating different capabilities of the A2A Agent Development Kit (ADK).

Table of Contents

• 📁 Structure
• 🚀 Quick Start
• 📚 Available Examples
• 🔧 Configuration
• 📖 Learning Path

📁 Structure

Each example is a self-contained scenario with:

• Server: A2A server implementation with task handlers
• Client: A2A client that demonstrates sending tasks and receiving responses
• Configuration: Environment-based config
• README: Detailed documentation and usage instructions

examples/
├── minimal/              # Basic server/client without AI (echo responses)
├── default-handlers/     # Using built-in default task handlers
├── static-agent-card/    # Loading agent config from JSON file
├── ai-powered/           # Server with LLM integration
├── ai-powered-streaming/ # AI with real-time streaming
├── streaming/            # Real-time streaming responses
├── artifacts-filesystem/ # Artifact storage using local filesystem
└── artifacts-minio/      # Artifact storage using MinIO (S3-compatible)

🚀 Quick Start

Running Any Example

1. Navigate to the example directory:

cd examples/minimal

1. Run with Docker Compose:

docker-compose up --build

1. Or run locally:

# Terminal 1 - Server
cd server && go run main.go

# Terminal 2 - Client
cd client && go run main.go

📚 Available Examples

Core Examples

minimal/

The simplest A2A server and client setup with custom echo task handler.

• Custom TaskHandler implementation
• Basic request/response pattern
• No external dependencies

default-handlers/

Server using built-in default task handlers - no need to implement custom handlers.

• WithDefaultTaskHandlers() for quick setup
• Automatic mock responses (no LLM required)
• Optional AI integration when LLM is configured

static-agent-card/

Demonstrates loading agent configuration from JSON files using WithAgentCardFromFile().

• Agent metadata defined in agent-card.json
• Runtime field overrides (URLs, ports)
• Environment-specific configurations

ai-powered/

Custom AI task handler with LLM integration (OpenAI, Anthropic, etc.).

• Custom AITaskHandler implementation
• Multiple provider support
• Environment-based LLM configuration

streaming/

Real-time streaming responses for chat-like experiences.

• Custom StreamableTaskHandler implementation
• Character-by-character streaming
• Event-based communication

ai-powered-streaming/

AI-powered streaming with LLM integration.

• Real-time AI responses
• Streaming LLM integration
• Event-driven architecture

Artifact Examples

artifacts-filesystem/

Demonstrates artifact creation and download using local filesystem storage.

• Filesystem storage provider
• HTTP download endpoints
• Client artifact download integration

artifacts-minio/

Demonstrates artifact creation and download using MinIO (S3-compatible) storage.

• MinIO storage provider
• S3-compatible API
• Production-ready cloud storage

🔧 Configuration

All examples follow a consistent environment variable pattern with the A2A_ prefix:

Common A2A Variables

• A2A_SERVER_PORT: Server port (default: 8080)
• A2A_DEBUG: Enable debug logging (default: false)
• A2A_AGENT_NAME: Agent identifier
• A2A_AGENT_DESCRIPTION: Agent description
• A2A_AGENT_VERSION: Agent version

AI/LLM Configuration

For examples with AI integration:

• A2A_AGENT_CLIENT_PROVIDER: LLM provider (openai, anthropic)
• A2A_AGENT_CLIENT_MODEL: Model to use (gpt-4, claude-3-haiku-20240307)
• A2A_AGENT_CLIENT_BASE_URL: Custom gateway URL (optional)

See each example's README for specific configuration details.

📖 Learning Path

Recommended progression:

1. minimal/ - Understand basic A2A protocol and custom task handlers
2. default-handlers/ - Learn built-in handlers for rapid development
3. static-agent-card/ - Externalize agent configuration to JSON files
4. artifacts-filesystem/ - Add file generation and download capabilities
5. ai-powered/ - Add LLM integration for intelligent responses
6. streaming/ - Implement real-time streaming capabilities
7. ai-powered-streaming/ - Combine AI integration with real-time streaming
8. artifacts-minio/ - Production-ready artifact storage with MinIO

For detailed setup instructions, configuration options, and troubleshooting, see each example's individual README file.

For more information about the A2A protocol and framework, see the main README or refer to the official documentation.