ComfyUI MGnodes Collection

A collection of custom nodes for ComfyUI focused on simplicity and usability.

Available Nodes

1. Image Watermark Node

Add image watermarks with advanced controls for transparency, positioning, and color manipulation. Features:

• Add image watermarks with adjustable transparency
• Multiple positioning options (Center, Top Left, Top Right, Bottom Left, Bottom Right, Tiled)
• Scale watermark size
• Remove white background from watermark
• Invert watermark colors
• Preserve transparency in both base image and watermark

2. Text Extractor Node

A utility node for processing text output from Language Models (LLMs), specifically handling Chain-of-Thought content.

Features:

• Extracts content between <think>...</think> tags
• Outputs both cleaned text and extracted thoughts
• Useful for processing LLM outputs with reasoning steps
• Supports multiple think blocks in a single input

Installation

Option 1: ComfyUI Manager (Recommended)

The easiest way to install is through the ComfyUI Manager. Simply search for "MGnodes" and click install.

Option 2: Manual Installation

1. Navigate to your ComfyUI custom nodes directory:

cd ComfyUI/custom_nodes/

2. Clone this repository:

git clone https://github.com/meanin2/comfyui-MGnodes.git

3. Restart ComfyUI

Usage

Image Watermark Node

1. Load your base image using a "Load Image" node
2. Load your watermark image using another "Load Image" node
3. Connect both to the "Add Image Watermark" node
4. Configure the watermark settings:
   • Position: Choose where to place the watermark
   • Opacity: Adjust transparency (0-100%)
   • Scale: Resize the watermark (10-1000%)
   • Make Black: Invert the watermark colors
   • Remove White: Make white pixels transparent

Text Extractor Node

1. Connect any text output (typically from an LLM node) to the Text Extractor
2. The node will output two strings:
   • Answer Text: The input text with all <think>...</think> sections removed
   • CoT Content: The extracted content from within the <think> tags

Example Workflow

Check out the examples folder for:

• A complete watermark workflow you can import into ComfyUI
• Sample images used in the showcase
• Test watermark images

Tips

• For best watermark results with "Remove White", ensure your watermark has a clean white background
• When using "Make Black", the inversion happens before white removal
• The watermark mask input is optional and can be used for additional transparency control
• Scale values above 100% will enlarge the watermark, below 100% will shrink it
• For text extraction, multiple <think> blocks will be joined with newlines in the CoT output