a2a mcp

Local 2025-09-01 01:04:53 0
Communication @khulnasoft-com/a2a-mcp

A centralized server that tracks and manages connected agents, providing a web interface to monitor their status while enabling agent communication through a central point.

This project demonstrates two different approaches to agent communication: 1. Master Control Program (MCP) - A centralized server-based approach where agents communicate through a central server 2. Agent-to-Agent (A2A) - A decentralized peer-to-peer approach where agents communicate directly with each other

Installation

  1. Create and activate a virtual environment: 

    python -m venv venv
source venv/bin/activate  # On Windows: venvScriptsactivate

  2. Install dependencies: 

    pip install -r requirements.txt

Usage

MCP Server and Agents

  1. Start the MCP server: 

    python cli.py run-mcp-server

  2. In separate terminals, start one or more MCP agents: 

    python cli.py run-mcp-agent --agent-id agent1
python cli.py run-mcp-agent --agent-id agent2

The MCP server will track all connected agents and their status. You can view the status by opening http://localhost:5000 in your browser.

A2A (Agent-to-Agent) Network

  1. Start the first A2A agent: 

    python cli.py run-a2a-agent --agent-id a2a1 --port 5001

  2. Start additional A2A agents, connecting them to existing agents: 

    python cli.py run-a2a-agent --agent-id a2a2 --port 5002 --peer localhost:5001
python cli.py run-a2a-agent --agent-id a2a3 --port 5003 --peer localhost:5001 --peer localhost:5002

A2A agents will automatically discover other agents through their initial peers. You can type messages in any agent s terminal to broadcast them to all connected agents.

Architecture

MCP (Master Control Program)

  • Centralized server that tracks all agents
  • Agents register with the server and maintain connection through heartbeats
  • Server provides a web interface to monitor agent status
  • Simple and reliable but has a single point of failure

A2A (Agent-to-Agent)

  • Decentralized peer-to-peer network
  • Agents connect directly to each other
  • Messages are flooded through the network
  • More resilient but requires more complex coordination
  • No single point of failure

Project Structure

a2a_mcp/
├── agents/              # Agent implementations
│   ├── mcp_agent.py    # MCP-based agent
│   └── a2a_agent.py    # Peer-to-peer agent
├── mcp/                # MCP server implementation
│   └── server.py       # Flask-based MCP server
├── cli.py             # Command-line interface
└── requirements.txt   # Python dependencies

Contributing

Feel free to submit issues and pull requests to improve the demonstration.