Empty image

This is the least glamorous node in the LF pack and one of the more useful ones: LF_EmptyImage generates a solid-color image tensor at whatever dimensions you ask for. That's it. No model, no AI, no download - three inputs, one output, done in milliseconds.

So why would you want a blank image? Because a shocking number of ComfyUI workflows need a canvas to start from. Img2img pipelines sometimes want a plain base to composite onto, then blend a generated region over it. Color-correction or grading chains want to test a filter against a known flat field. Mask work wants a neutral canvas to build on. And when you're debugging - "is my resize node broken, or my image?" - generating a known solid-color tensor isolates the problem instantly. A predictable blank is worth more than it sounds.

Inputs and outputs

• width - default 1024, range 64–8192, steps of 64.
• height - default 512, range 1–4096 (oddly, step 1, so you can make non-multiple heights if you must).
• color - a hex string in RRGGBB format (no #), default 000000 (black). FFFFFF for white, any color you like.

Single output: image, the generated tensor. Feed it into anything that accepts an IMAGE.

Installing it

One pack, as with every node here:

cd ComfyUI/custom_nodes
git clone https://github.com/lucafoscili/lf-nodes

restart ComfyUI, or install "LF Nodes" via ComfyUI Manager. Zero dependencies beyond what the suite already loads - this is pure tensor construction.

Notes

• Don't confuse it with the Empty Latent Image node in core ComfyUI. That one creates a latent (used by the sampler), this one creates an actual visible pixel image. Different animals.
• Remember the RRGGBB format has no # prefix - it's a small gotcha when you're copying hex codes from elsewhere. The node is strict about it.
• If you find yourself generating an empty image more than "occasionally," you're probably building a compositing pipeline, and you've just found the canvas generator for it.