Skip to content

Multi-Agent A2A Communication Example

Demonstrates specialized A2A agents discovering and communicating with each other.

Quick Start

# Install dependencies
pip install -r requirements.txt

# Set environment variable (optional for AI responses)
export GEMINI_API_KEY="your-gemini-api-key"

# Run the multi-agent demo
python main.py

What This Demo Shows

  1. 3 Specialized Agents start on different ports:
  2. DataAnalyst (port 8001): Data analysis and insights
  3. ContentWriter (port 8002): Content creation and marketing

  4. CodeReviewer (port 8003): Code review and engineering

  5. A2A Discovery: Each agent discovers others via /.well-known/agent.json

  6. Agent-to-Agent Communication:

  7. DataAnalyst analyzes sales data
  8. ContentWriter creates marketing content based on the analysis

A2A Protocol Compliance

Each agent implements:

  • ✅ Agent Discovery: /.well-known/agent.json endpoint
  • ✅ Task Processing: /tasks/send endpoint
  • ✅ Specialized Capabilities: Unique skills and descriptions
  • ✅ A2A Message Format: Proper request/response structure

Example Workflow

# 1. DataAnalyst analyzes data
curl -X POST http://localhost:8001/tasks/send \
  -H "Content-Type: application/json" \
  -d '{
    "id": "task-123",
    "message": {
      "role": "user",
      "parts": [{"type": "text", "text": "Sales increased 25% last quarter"}]
    }
  }'

# 2. ContentWriter creates content based on analysis
curl -X POST http://localhost:8002/tasks/send \
  -H "Content-Type: application/json" \
  -d '{
    "id": "task-456",
    "message": {
      "role": "user",
      "parts": [{"type": "text", "text": "Create marketing content for 25% sales growth"}]
    }
  }'

Agent Endpoints

While running, each agent exposes:

  • DataAnalyst: http://localhost:8001
  • ContentWriter: http://localhost:8002
  • CodeReviewer: http://localhost:8003

Each with: - /.well-known/agent.json (A2A discovery) - /tasks/send (A2A task processing) - /health (health check)

Features

  • 🤖 Specialized Agents with unique capabilities
  • 🔍 Automatic Discovery via A2A protocol
  • 💬 Inter-Agent Communication with task delegation
  • 🧠 AI Integration Example via Gemini (optional)
  • 📊 Real Workflow demonstration (Data Analysis → Content Creation)

This example is based on the working tests/test_e2e_multi_agent.py and demonstrates real A2A Protocol multi-agent collaboration!