GetAndon
Back to Blog

The MCP Protocol: How Andon Communicates with AI Assistants

A
Alex Chen, CTO
May 5, 2025 • 3 min read

The MCP Protocol: How Andon Communicates with AI Assistants

The Model Communication Protocol (MCP) is the backbone of Andon's ability to communicate with various AI assistants. In this post, we'll explore what MCP is, how it works, and why it's essential for creating a unified status monitoring system.

What is the MCP Protocol?

MCP is a standardized communication protocol designed specifically for AI models and assistants to report their status, progress, and needs to external devices and applications. It provides a consistent way for different AI systems to share their operational state, regardless of their underlying architecture.

Key Components of MCP

The protocol consists of three main components:

1. Status Reporting

AI assistants continuously broadcast their current status:

```json { "status": "working", "progress": 0.65, "estimated_completion_time": "2025-05-05T15:30:45Z", "model_id": "cursor-v2" } ```

2. Event Notifications

When significant events occur, the AI sends specific notifications:

```json { "event_type": "attention_required", "message": "Additional information needed to complete task", "severity": "medium", "timestamp": "2025-05-05T15:28:30Z" } ```

3. Command Interface

External devices can send commands to the AI:

```json { "command": "pause_processing", "parameters": { "save_state": true } } ```

How Andon Uses MCP

Andon listens for MCP broadcasts from compatible AI assistants and translates them into visual cues:

  1. Status Monitoring: Andon continuously monitors the status reports from all connected AI assistants.

  2. Color Mapping: Each status is mapped to a specific color:

    • Idle → Green (or your custom color)
    • Working → Blue (or your custom color)
    • Attention Required → Red (or your custom color)

  3. Priority System: When multiple AI assistants are connected, Andon uses a priority system to determine which status to display:

    • Attention Required (highest priority)
    • Working (medium priority)
    • Idle (lowest priority)

Implementing MCP Support

For developers interested in adding MCP support to their own AI assistants or applications, the process is straightforward:

  1. Install the MCP Library: ```bash npm install mcp-protocol ```

  2. Initialize the MCP Client: ```javascript import { MCPClient } from 'mcp-protocol';

    const mcp = new MCPClient({ application_name: 'my-ai-assistant', version: '1.0.0' }); ```

  3. Send Status Updates: ```javascript // When AI starts processing mcp.updateStatus('working', { progress: 0.0 });

    // During processing mcp.updateStatus('working', { progress: 0.5 });

    // When complete mcp.updateStatus('idle');

    // When user input is needed mcp.updateStatus('attention_required', { message: 'Please provide additional context' }); ```

The Future of MCP

We're actively working with AI assistant developers to expand MCP adoption and enhance its capabilities. Future versions of the protocol will include:

  • More granular status reporting
  • Customizable event types
  • Bidirectional communication for advanced control
  • Enhanced security features

Conclusion

The MCP protocol is what makes Andon a universal solution for monitoring AI assistants. By providing a standardized way for AI systems to communicate their status, MCP enables seamless integration between Andon and a wide range of AI tools.

If you're developing an AI assistant and would like to add MCP support, please reach out to our developer relations team at [email protected].

Alex Chen, CTO