How marioluciofjr/mapas_mentais_mcp works
Project type and purpose
• This is a Python‐based MCP-Server tool (“mapas_mentais”) that acts as a parent MCP server. It dynamically launches and manages child MCP servers (one per mental-map operation) and integrates directly with Claude Desktop via the Model Context Protocol. Its core functionality is to generate six types of automated mental maps by invoking Anthropic’s Claude models through the FastMCP SDK and the uv command-line wrapper.
File structure
Top-level files only—no nested “frontend” or “backend” folders—indicate a single‐process, CLI-driven microservice:
• pyproject.toml + uv.lock → build and dependency management via UV (Astral)
• main.py → likely the script to bootstrap or delegate to server.py
• server.py → the MCP server implementation registering six mental map handlers
• README.md, LICENSE, .gitignore, .python-version → supporting and configuration files
There are no separate “lib” or “tests” directories, so the architecture revolves around one runtime component.
README for additional insights
• Uses Python ≥3.10, the uv package, and mcp[cli] (Anthropic’s Python SDK)
• Describes six commands (“apresenta”, “compara”, etc.) registered with FastMCP
• Integrates with Claude Desktop by adding an entry in claude_desktop_config.json that calls uv run ... server.py
• Runtime flow: user selects a tool in Claude Desktop → UV spawns the MCP server → commands get dispatched → each handler calls the Claude model → returns a structured mental map
Final recommendations
• Deployment context: annotate that everything runs locally under the .venv and is launched by Claude Desktop
• Validate with code: cross-check the imports in server.py and main.py to ensure you’ve included all critical libraries and flows