Nuke-MCP

A bridge between The Foundry's Nuke and AI systems using the Model Context Protocol (MCP).

Overview

Nuke-MCP allows AI assistants to interact with Nuke through a socket connection, enabling them to:

• Get information about Nuke scripts
• Create, modify, and delete nodes
• Position nodes in the node graph
• Connect nodes together
• Control playback and rendering
• Execute arbitrary Python code in Nuke

Components

1. Nuke Addon (nuke_mcp_addon.py): A Nuke script that creates a socket server within Nuke
2. MCP Server (nuke_mcp_server.py): A Python server that connects to the Nuke addon and exposes tools to AI systems
3. Entry Point (main.py): A simple script to start the MCP server

Installation

Prerequisites

• The Foundry's Nuke (any recent version should work)
• Python 3.7+
• FastMCP package
• Claude Desktop

Installing the FastMCP package

pip install fastmcp

Installing Nuke-MCP

1. Clone or download this repository
2. Copy nuke_mcp_addon.py to your Nuke scripts folder or a location in your Nuke Python path

Usage

Installing the Nuke Addon

1. Copy the Addon File:

   • Take the nuke_mcp_addon.py file and place it in a location where Nuke can find it:
     • Copy it to your Nuke scripts folder (usually in ~/.nuke/python on Linux/Mac or in your home directory on Windows)
     • Or place it in a folder that's in your Nuke Python path

2. Create a Startup Script (Recommended):

   • Create or edit an existing init.py file in your .nuke directory
   • Add the following line to automatically load the addon when Nuke starts and making sure you're pointing to the python subfolder where the MCP tools live:

     nuke.pluginAddPath("./python")
     
     import nuke_mcp_addon
     

3. Manual Loading (Alternative):

   • If you don't want to load it automatically, you can manually load it each time by:
     • Opening Nuke
     • Make sure you have a python subfolder setup in the init.py where the MCP tools live
     • Going to the Script Editor panel
     • Running import nuke_mcp_addon

Docking the NukeMCP Panel

By default, the NukeMCP panel opens as a floating window. If you prefer to have it docked in Nuke's interface, you can modify the NukeMCPPanel class in nuke_mcp_addon.py:

# Find the NukeMCPPanel class definition (around line 380)
class NukeMCPPanel(nukescripts.PythonPanel):
    def __init__(self):
        nukescripts.PythonPanel.__init__(self, 'Nuke MCP', 'com.example.NukeMCP')
        # ... existing code ...

# Add this method to enable docking
    def addToPane(self):
        pane = nuke.getPaneFor('Properties.1')
        if not pane:
            pane = nuke.getPaneFor('Viewer.1')
        self.setMinimumSize(300, 200)  # Set a reasonable minimum size
        return pane.addPermanentAsQWidget(self)

# Modify the show_panel function to use docking
def show_panel():
    """Show the NukeMCP panel"""
    global _panel
    if _panel is None:
        _panel = NukeMCPPanel()
    
    # Show as docked panel instead of floating window
    pane = _panel.addToPane()
    if pane:
        _panel.setParent(pane)

You can also modify how the panel appears in the menu. Based on your menu.py file, you have:

# MCP Tools
nuke.toolbar("Nodes").addCommand('NukeMCP/NukeMCP Panel', 'nuke_mcp_addon.show_panel()')

To ensure it's properly integrated with your existing toolbar structure, make sure this line is uncommented in your menu.py file.

Troubleshooting

To use Nuke-MCP with Claude Desktop, follow these steps:

1. Download and Install Claude Desktop:

   • Download Claude Desktop from Anthropic's website
   • Install and set up your account

2. Enable Developer Mode:

   • Open Claude Desktop
   • Go to Settings
   • Enable Developer Mode (this might require specific permissions)

3. Edit the Configuration File:

   • In Claude Desktop settings, click "Edit Config" (or directly ed

... View full README on GitHub