Black Image Generator

Black Image Generator does one thing: it creates a solid black RGB image of whatever size you ask for. No prompt, no diffusion, no VAE - it's a torch.zeros call in a trench coat. You will never be impressed by it. You may, however, find yourself reaching for it constantly, because ComfyUI workflows are full of IMAGE inputs that you sometimes just need to fill.

When you'd use it

The honest list of use cases is short but real:

• A neutral placeholder so a workflow stays runnable while you swap in a real input.
• A starting canvas to composite other images onto, or to mask against.
• An img2img seed image when you genuinely want the model to start from black (some restoration/denoise approaches like it).
• Feeding a batch of blank frames into a node that requires an IMAGE before its real source is wired up.

It's the utility equivalent of a blank sheet of paper: most useful in the middle of a build, not as a destination.

The inputs

• width (INT, 64–8192, step 8, default 512)
• height (INT, 64–8192, step 8, default 512)
• batch_size (INT, 1–64, default 1)

Output is a single image IMAGE tensor of shape [batch, height, width, 3], values in the 0–1 float range ComfyUI expects. All zeros, as promised.

Gotchas

The values are pure black (0.0) in every channel. If you were hoping for "dark gray" or transparency, that's not this node - black means black, and there's no alpha channel. If you need it as a mask, remember a mask is 0–1 too, so a fully black image is a fully zero mask, which is "no mask" to most nodes. You'll want to invert or threshold it elsewhere if that's the plan.

Also note the 8-pixel step on width/height. It'll happily hand you something like 512x520 if you spin a slider - fine for image work, but if that black canvas is destined to become a latent, you may want multiples of 16, which is what the pack's Image Dimensions (Multiple of 16) node is for.

Installation

From Praveen's ComfyUI Tools, a small solo pack. Install via ComfyUI Manager (search "Praveen" / "praveen-tools"), or:

cd ComfyUI/custom_nodes
git clone https://github.com/Praveenhalder/praveen-tools

Restart ComfyUI. Zero extra dependencies - the pack's requirements are just Pillow, NumPy, and PyTorch, all bundled with a stock ComfyUI install. No models, no downloads.

This is the kind of node that gets zero Google impressions because nobody searches for "I need a black rectangle." But the pack's README lists it for a reason: utility nodes like this one are what keep a graph from stalling on a missing input.