Image Select Color

Per WAS's own description, this one does something narrow and literal: it returns your selected color, and only that color, sitting on an otherwise black canvas. Anything in the image that isn't a match for the target color gets replaced with black; anything that is stays.

What it's actually for

This is a blunt instrument, not smart segmentation - no object detection, no semantic understanding, just literal color matching. That makes it a poor fit for isolating "the subject" in a photograph, where lighting and anti-aliasing mean the same object is made of thousands of subtly different pixel values. Where it does work well is flat, limited-palette content: a mockup with a solid-color placeholder region you want to isolate, pixel art, a UI screenshot where a specific element uses one consistent color, or a color-coded diagram where regions are keyed by hue. Point it at a specific color and it hands you back a mask-like image showing exactly, and only, where that color appears.

It's also a quick debugging tool. If you're trying to figure out where pure white, pure black, or a specific brand color ended up in a generated image, this node answers that visually faster than eyeballing the full image.

How it works

You feed it an image and tell it which color to look for; how forgiving the match is against near-identical shades is presumably tunable, since exact-pixel-only matching would make the node nearly useless on anything but flat vector-style art. The output is a same-size image: black canvas, with only the matched pixels showing through in their original color.

Installing it

Comes with the full WAS Node Suite pack:

• ComfyUI Manager - search WAS Node Suite, install, restart.
• Manual - cd ComfyUI/custom_nodes && git clone https://github.com/WASasquatch/was-node-suite-comfyui/, then pip install -r requirements.txt from inside that folder, then restart.

Common issues

The most common disappointment with this node is running it on a photograph and getting either almost nothing selected or a scattered, noisy result - that's a symptom of exact color matching meeting a real photo, where "the same" color varies pixel to pixel because of lighting, compression, and anti-aliasing. If you loosen the match tolerance and it's still not behaving, sample the exact color you're targeting straight from the source image (with a color picker) rather than typing in a value from memory - a small mismatch there is enough to break a strict match entirely.

Beyond that, the usual WAS caveat applies: the pack has been unmaintained since December 2023, and "Import Failed" errors after a ComfyUI update are a known, recurring issue tied to the pack's older pinned dependencies colliding with newer ones. Reinstalling requirements.txt in your ComfyUI venv is the standard fix if the whole suite stops loading.