BMAB Black And White

Sometimes you don't need the fanciest tool in the pack - you need a node that takes an image and turns it grayscale, full stop. That's this one.

What it is and why you'd reach for it

BMAB Black And White converts a color image to grayscale. There's no threshold to tune, no channel weighting to pick, no options at all beyond the image itself. That simplicity is the whole point: it's a utility node for the moments where you need a quick desaturation step inside a larger pipeline, rather than a standalone editing tool.

Where this actually shows up in practice: prepping an image as an edge or line-art condition for ControlNet (many preprocessors want or produce grayscale-adjacent input), building a black-and-white stylistic pass, feeding a grayscale image into a compositing step where you want luminance without color, or just checking how an image reads structurally without color influencing your eye. It's also a handy sanity-check node when you're debugging a graph - grayscale output makes it obvious at a glance whether a downstream node is receiving the image you think it is.

How it works

Standard image-to-grayscale conversion - luminance is computed from the color channels and applied uniformly across R, G, and B, so the output is still a 3-channel IMAGE (not a single-channel mask), it just no longer carries any color information. That matters if a downstream node specifically expects a MASK type rather than an IMAGE - this node stays in IMAGE-land throughout.

Inputs and outputs

As minimal as it gets:

• image (required) - the color image to convert.

Output: a single image, grayscale, same dimensions as the input.

Installing it

Through ComfyUI Manager: search comfyui_bmab, install, restart ComfyUI. Manual path:

cd ComfyUI/custom_nodes
git clone https://github.com/portu-sim/comfyui_bmab.git
cd comfyui_bmab
pip install -r requirements.txt

(Windows portable: use python_embeded\python.exe -m pip install -r requirements.txt instead.) No extra dependencies beyond BMAB's own base requirements - this node doesn't need comfyui_controlnet_aux or ComfyUI_IPAdapter_plus, the two optional packs BMAB's ControlNet-related nodes rely on.

Common issues

Honestly, there isn't much to go wrong with a single-input grayscale conversion. The only things worth flagging:

• Expecting a MASK, getting an IMAGE. If you're chaining this into a node that specifically wants a mask type, you'll need an extra conversion step - this node's output stays an image, just a desaturated one.
• Looking for adjustable grayscale weighting and not finding it. If you need control over how much each color channel contributes to the grayscale result (for a specific stylistic look, say), this node doesn't expose that - it's fixed conversion. You'd want a more general color-adjustment node for that level of control.