a2a-demo

# A2A Demo A demonstration project showcasing Agent-to-Agent (A2A) communication using the Strands framework. ## Overview This project demonstrates how to create and deploy AI agents that can communicate with each other using the A2A protocol. It includes a basic greeting agent that provides personalized greetings with fortune messages. ## Project Structure ``` a2a-demo/ ├── main.py # Main entry point ├── pyproject.toml # Project configuration ├── demos/ │ └── basic_agent/ # Basic greeting agent demo │ ├── main.py # Agent implementation │ ├── pyproject.toml # Agent-specific config │ └── README.md # Agent documentation └── README.md # This file ``` ## Features - **Basic Greeting Agent**: An A2A-enabled agent that provides personalized greetings with fortune messages - **A2A Server**: HTTP server for agent-to-agent communication - **Configurable Port**: Environment variable support for server port configuration ## Requirements - Python >= 3.14 - Dependencies managed via `uv` ## Installation 1. Clone the repository 2. Install dependencies: ```bash uv sync ``` ## Quick Start 1. **Install dependencies:** ```bash uv sync ``` 2. **Start the server:** ```bash cd demos/basic_agent uv run main.py ``` 3. **Retrieve the A2A AgentCard:** ```bash curl http://localhost:8000/.well-known/agent.json ``` The server runs on port 8000 by default. Set `A2A_SERVER_PORT` to use a different port. ## Usage ### Running the Basic Agent Navigate to the basic agent demo and start the server: ```bash cd demos/basic_agent uv run main.py ``` The agent will start an A2A server on port 8000 (default) or the port specified by the `A2A_SERVER_PORT` environment variable. ### Environment Variables - `A2A_SERVER_PORT`: Port for the A2A server (default: 8000) ## Agent Tools The basic agent provides the following tool: - **greet(name: str)**: Generates a personalized greeting with a random fortune message ## Development This project uses a workspace structure with `uv` for dependency management. The main project coordinates multiple agent demos, each with their own dependencies and configurations. To add new agent demos: 1. Create a new directory under `demos/` 2. Add the new demo to the workspace members in the root `pyproject.toml` 3. Implement your agent using the Strands framework ## Dependencies - `strands-agents[a2a]`: Core agent framework with A2A support - `fortune-python`: Fortune message generation