Preview Mask Bridge (Yogurt Nodes)

Masks are the thing you least want to trust blind. A slightly wrong mask means an inpaint or a ControlNet pass quietly doing the wrong thing, and the built-in way to check is to drag a preview node onto the branch and kill the flow. YogurtPreviewMaskBridge is the pass-through version: it shows you the mask and hands it right back, so you can look before you commit compute.

• masks in - the MASK tensor.
• masks out - the same tensor, untouched.
• width, height, batch out - the mask's dimensions and count.

It's a preview node and a free metadata source at once. The classic spot for it is right after mask generation or mask-drawing, before the mask feeds an inpaint sampler or a detailer. You get to see exactly what the mask captured - and because it passes through, you don't have to rebuild the branch just to look at it.

Why you want the dims, too

Masks that don't match their image's resolution are a classic source of "why is my inpaint garbage" bugs. width and height here let you compare against the image's real dimensions without guessing. Feed them into whatever expects the shape, and you'll catch the mismatch at the source instead of after a full sampling pass.

Install

Ships in ComfyUI-YogurtNodes, same as the image and "any" bridges. ComfyUI Manager → search ComfyUI-YogurtNodes → install, or:

cd ComfyUI/custom_nodes
git clone https://github.com/yogurt7771/ComfyUI-YogurtNodes.git
cd ComfyUI-YogurtNodes
pip install -r requirements.txt

Restart ComfyUI, find it under "Yogurt Nodes" with the " (Yogurt Nodes)" suffix on the name. No models, no keys.

One note: masks render as grayscale previews, so if two masks look identical you can still be looking at different values - use the batch output or a proper mask-stat node if you need the numbers, not just the picture. For a quick visual sanity check, though, this is the node.