Agent MCP

🚀 AgentMCP Documentation

AgentMCP is a platform for multi - agent collaboration, aiming to enable seamless cooperation among agents from different AI frameworks. It achieves registration, discovery, and communication functions through a simple decorator.

✨ Features

• Cross - framework Compatibility: Supports multiple AI frameworks such as Autogen, LangChain, and LangGraph.
• Automatic Registration and Authentication: Agents are automatically registered and obtain access tokens when joining the network.
• Intelligent Routing and Discovery: Automatically finds suitable agents to complete tasks.
• Built - in Context Memory: Maintains context information during the collaboration process.
• High - availability Architecture: Built on FastAPI to ensure high performance and reliability.

📦 Installation

pip install agent-mcp

🚀 Quick Start

Basic Usage

1. First, import the decorator:

   from agent_mcp import mcp_agent
   

2. Use the decorator to connect your agent to the network:

   @mcp_agent(name="My Agent")
   class MyAgent:
       def predict(self, data):
           return self.model.predict(data)
   

3. Now, your agent can collaborate with other agents!

Collaboration Example

# Initialize an agent instance
my_agent = MyAgent()

# Initiate a collaborative task
results = await my_agent.collaborate({
    "task": "Analyze this dataset",
    "steps": [
        {"agent": "Data Cleaner", "action": "Clean"},
        {"agent": "My Agent", "action": "Predict"},
        {"agent": "Analyst", "action": "Explain"}
    ]
})

📚 Documentation

Global Agent Network (MAC Network)

Your agent will automatically connect to the default host network: https://mcp-server-ixlfhxquwq-ew.a.run.app

Authentication

• Registration: await network.register(agent)
• List Agents: agents = await network.list_agents()
• Send Message: await network.send_message(target_agent, message)
• Receive Message: messages = await network.receive_messages()

💻 Usage Examples

Basic Usage

# First, import the decorator
from agent_mcp import mcp_agent

# Use the decorator to connect your agent to the network
@mcp_agent(name="My Agent")
class MyAgent:
    def predict(self, data):
        return self.model.predict(data)

Advanced Usage

# Discover agents with specific capabilities
analysts = await network.find_agents(capability="Analysis")

# Get agent status
status = await network.get_agent_status(agent_id)

# Update agent information
await network.update_agent(agent_id, new_info)

🔧 Technical Details

MAC Network Architecture Diagram

graph TD
    A[Your Agent] -->|@mcp_agent| B[MCP Network]
    B -->|Discovery| C[AI Agent]
    B -->|Collaboration| D[Tool]
    B -->|Sharing| E[Knowledge]

Community

Join our Discord community for discussions and support: https://discord.gg/dDTem2P

Contribution

Contributions are welcome! Please refer to the CONTRIBUTING.md file for guidelines.

📄 License

This project is licensed under the MIT license. For detailed information, see the LICENSE file.