Extensions/ComfyUI-bby-nodes
ComfyUI Extension

ComfyUI-bby-nodes

Custom utility nodes for ComfyUI

By Eonizer·Created 8 months ago·Updated a day ago· 0
Eonizer/ComfyUI-bby-nodes
Nodes4
On cloudLocal install
CategorybbyNodes
Stars0
Updateda day ago
Readme

ComfyUI-bby-nodes

A custom node pack for ComfyUI.

Installation

Clone into your ComfyUI custom_nodes directory:

cd custom_nodes
git clone https://github.com/Eonizer/ComfyUI-bby-nodes

Restart ComfyUI.

Nodes

TagFilterNode

Filters a comma/newline-separated list of tags against a reference tag list.

Outputs:

  • filtered_tags — tags that matched the reference list
  • missing_tags — tags from the reference list not found in the input

Options:

  • Match mode: exact or partial
  • Case sensitivity toggle
  • Underscore-stripping

Concept Picker

Code picks the concepts, the LLM only writes the prompt.

  • Concept List — one per category (subject, setting, style…); chain them together via the concepts socket. Picks between min_count and max_count items (the seed picks the count) from a newline-separated list (# lines are comments).
    • Seed control: fixed = same pick every run, randomize = new pick every run, increment = walks the whole list with no repeats.
    • pin — type a value to force it, ignoring the list.
    • file — load the list from a .txt file in the pack's lists/ folder (subfolders allowed) instead of the items box. Press R in ComfyUI to refresh the dropdown after adding files; editing a selected file re-runs the node.
    • file_path — or type a path to a .txt list anywhere the ComfyUI machine can read, e.g. D:\lists\subjects.txt or a network share \\server\share\subjects.txt (surrounding quotes, ~ and %VARS% are fine; relative paths start in lists/). Overrides file.
  • Concept Prompt Builder — turns the chain into user_prompt + system_prompt strings for any LLM node (e.g. "OpenAI API - Chat Completion"). extra_details limits how many details the model may add; {extra} in the system prompt is replaced with that rule.
  • Aspect Ratio From Text — reads an [AR:2:3] tag from the LLM reply and outputs the nearest "Resolution (CRT)" dropdown value, the reply with the tag removed, and width/height for the given megapixels.