guiMCP provides a blueprint-style visual node editor for building FastMCP servers without writing code. You can define MCP server components such as Tools, Resources, and Prompts by wiring together execution and data nodes, then preview generated Python code in real time. The editor supports live simulation of tool calls, so you can test individual nodes directly within the UI, and you can save projects as .guimcp files for later editing. When you’re ready, you can export clean, runnable FastMCP server code and run it with Python to start your server.

To use it, start the GUI, add a Server Config node to name your MCP server, then add Tool Definition, Resource, and Prompt nodes as needed. Build your logic with Flow Control and Data nodes, connect to a Return node to define the output, and use the Export Python feature to generate a mcp_server.py file. You can run the exported server with python mcp_server.py to boot the FastMCP server and test calls against your configured tools.

Prerequisites:

• Python 3.11+ installed on your system
• Pip available in your PATH
• Optional: PySide6 and NodeGraphQt for the GUI (as listed in requirements)

Installation steps:

1. Clone the repository git clone https://github.com/yourusername/guiMCP.git cd guiMCP

2. Create and activate a virtual environment (recommended) python -m venv venv

   On Windows

   venv\Scripts\activate

   On macOS/Linux

   source venv/bin/activate

3. Install dependencies pip install -r requirements.txt

4. Run the GUI python main.py

5. Optional: If you export a server, run the exported server code python mcp_server.py

Tips and notes:

• Ensure you have FastMCP installed or available in your Python environment when running exported servers.
• The GUI exports Python code that targets FastMCP; review the generated code to customize or extend functionality.
• If the editor cannot connect to certain libraries (e.g., PySide6 or NodeGraphQt), install them explicitly via pip and ensure your Python version matches the requirements.
• Save projects frequently as .guimcp files to avoid data loss, and use File → Export Python (Ctrl+E) to generate runnable server code.
• When testing, use the Simulate Tool Calls feature by right-clicking a tool node to validate behavior before exporting.
• If you encounter environment variable issues, you can set API keys, URLs, or server names via an optional env block in the mcp_config once your deployment requires it.