ComfyUI-Mircify

A ComfyUI custom node for converting images into IRC art by processing them in pixel blocks and generating both image and text outputs with optimal color selection from a 99-color IRC palette.

This node is used with aibird to convert images to IRC art.

Workflow

An example workflow is provided in workflows/mircify_example_workflow.json. At the moment this example is using qwen image lightning, however any model that is decent at making pixel ansi style art can suffice.

Features

• Universal Image Support: Works with any image dimensions
• Color Transfer Integration: Uses KMeans clustering with multiple distance methods for accurate color mapping
• Block Processing: Configurable block sizes (default 8x15 pixels)
• Half Block Mode: Option for 8x7.5 pixel blocks for finer resolution
• Dominant Color Extraction: Uses k-means clustering to find the most prominent color in each block
• Multiple Distance Methods: Choose between weighted manhattan, manhattan, euclidean, and perceptual weighted distance calculations
• 99-Color IRC Palette: Maps colors to extended IRC color palette
• 16-Color Compatibility Mode: Option to use traditional 16-color IRC palette
• Dual Output: Returns both processed image and mIRC-formatted text
• Optimized Text Output: Compressed color codes and space-efficient formatting
• Text File Export: Save IRC codes directly to text files with timestamp support

Installation

1. Clone or download this repository to your ComfyUI custom_nodes directory:

   cd ComfyUI/custom_nodes
   git clone https://github.com/birdneststream/ComfyUI-Mircify
   

2. Install dependencies:

   cd ComfyUI-Mircify
   pip install -r requirements.txt
   

3. Restart ComfyUI

Nodes

IRC Art Converter

Category: image/conversion

Converts images into IRC art with both visual and text output.

Inputs:

• image: Input image (any dimensions supported)
• block_width: Width of each block in pixels (default: 8)
• block_height: Height of each block in pixels (default: 15)
• half_block_mode: Enable half-height blocks (8x7.5 instead of 8x15)
• use_16_colors: Use only first 16 IRC colors for compatibility (default: False)
• distance_method: Color distance calculation method (weighted_manhattan/manhattan/euclidean/perceptual_weighted)

Outputs:

• processed_image: Visual representation of the IRC art conversion
• irc_text: mIRC-formatted text with color codes ready for IRC clients

IRC Text Saver (still not working too good with color matching)

Category: text/output

Saves IRC-formatted text to files for easy copying.

Inputs:

• irc_text: IRC text from the converter (STRING input)
• filename: Output filename with {TIMESTAMP} placeholder support (default: "irc_art.txt")

Outputs:

• file_path: Absolute path to the saved file

Usage

1. Basic Workflow:

   • Load an image
   • Connect it to "IRC Art Converter"
   • Connect the irc_text output to "IRC Text Saver"
   • Run the workflow

2. Distance Methods:

   • weighted_manhattan: Manhattan distance with perceptual weights (Green=1.0, Red=0.7, Blue=0.5) for better natural color matching
   • manhattan: Standard Manhattan distance (L1 norm) - fastest calculation
   • euclidean: Standard Euclidean distance (L2 norm) - geometric accuracy
   • perceptual_weighted: Euclidean distance with CIE luminance weights for perceived brightness matching

3. Block Modes:

   • Full blocks (8x15): Standard IRC character block size
   • Half blocks (8x7.5): Provides finer vertical resolution

4. Color Modes:

   • 99-color mode: Full extended IRC palette for maximum color accuracy
   • 16-color mode: Traditional IRC colors for better compatibility

5. Text Output:

   • Uses optimized mIRC color text output
   • Ready to paste directly into IRC clients
   • Support for {TIMESTAMP} placeholder in filenames

Color Processing

The node uses a two-stage approach:

1. Color Transfer Phase: Applies KMeans clustering with configurable distance methods to map the entire image to the 99-color IRC palette
2. Block Processing Phase: Divides the color-corrected image into blocks using dominant color extraction and generates the final IRC art output

IRC Color Palette

The node uses an extended 99-color IRC palette built in. It also can support just 16 colors.

For more information about how IRC uses colors refer to these resources:

• https://modern.ircdocs.horse/formatting.html#color
• https://anti.teamidiot.de/static/nei/*/extended_mirc_color_proposal.html

Requirements

• ComfyUI
• Python 3.8+
• torch
• numpy
• scikit-learn
• pillow

Credits

• Color transfer logic adapted from ComfyUI-Color_Transfer by 45uee
• Uses KMeans clustering with modified multiple distance methods for optimal color mapping

License

MIT License