MCP Agent Graph (MAG)

What is MCP Agent Graph (MAG)?

MCP Agent Graph (MAG) is an agent development framework for rapidly building multi‑agent systems using graphs, nodes, and the Model Context Protocol (MCP). It runs as a web application with a Python backend and a JavaScript frontend, targeting developers who need to design, visualize, and deploy complex agent workflows.

How to use MCP Agent Graph (MAG)?

Install the Python SDK via pip install mcp-agent-graph, then clone the repository. Start the backend with python main.py (on port 9999) and the frontend with npm run dev (on port 5173). For a quick start, import the provided deepresearch.zip file into the frontend canvas.

Key features of MCP Agent Graph (MAG)

• AI‑generated agents from natural language requirements
• Visual graph editor with drag‑and‑drop node configuration
• Each node functions as an independent agent with tool capabilities
• Graph nesting: embed entire graphs as nodes in higher‑level graphs
• One‑click export of graphs as standard MCP server scripts
• Complete agent packaging, sharing, and deployment lifecycle
• Python SDK for integrating visual designs with code execution

Use cases of MCP Agent Graph (MAG)

• Building a deep research system that performs multi‑round retrieval and generates HTML reports
• Simulating complex corporate ethical dilemmas with an AI CFO