imageConcat

Sometimes you just need two images stuck side by side. A before/after. A reference next to a result. Two halves you want the model to see as one canvas. easy imageConcat does exactly that and nothing more - it takes two images and joins them into one, in whichever direction you point it.

It's a plumbing node from yolain's ComfyUI-Easy-Use pack, and its value is entirely in saving you from doing this manually. No exporting, no image editor, no re-importing. Wire two images in, get one out, keep the whole thing inside your graph.

How it works

The node pastes image2 onto image1 along an edge. The direction input decides which edge: right puts image2 to the right of image1, down stacks it below, and left/up do the mirror. That's the core of it.

The one setting that saves you a headache is match_image_size. If your two images aren't the same height (for a left/right join) or width (for up/down), the concat can look ragged or fail to line up. Flip match_image_size on and the node resizes image2 to fit image1's matching dimension first, so the seam is clean. Leave it off only when you know the two already match and you want the pixels untouched.

Inputs and outputs

There are only four inputs and two of them are the images. The ones you'll actually touch:

• direction - right, down, left, or up.
• match_image_size - turn this on when the two images differ in size, which is most of the time.

The single output is IMAGE, the combined result. It wires into anything that takes an image: a preview, a save node, or straight back into a VAE encode if you want the model to treat the pair as one input.

Where you'd use it

The obvious case is comparison shots - pipe your source and your generated version in, get a clean side-by-side to post or save. The more interesting case is feeding stitched pairs into reference-driven workflows (IC-LoRA and similar setups expect a two-panel image where one side is the reference and the other is what you're generating). imageConcat is a tidy way to build that panel without leaving ComfyUI.

Installing it

Grab the whole pack: in ComfyUI Manager search "ComfyUI Easy Use," install, restart. Or cd ComfyUI/custom_nodes && git clone https://github.com/yolain/ComfyUI-Easy-Use, run install.bat (or pip install -r requirements.txt), and restart. First boot may switch your theme to the pack's Obsidian UI - Settings → Color Palette to change it back.

Common issues

The classic gotcha is a mismatched-size join producing weird black bars or an off-looking seam - that's match_image_size being off when it should be on. Turn it on.

And don't confuse this with easy joinImageBatch. imageConcat takes two specific images on named inputs. joinImageBatch takes a whole batch and lays it out as a strip. If you're trying to combine three or more, or the output of a batch generation, that's the other node - imageConcat only knows about image1 and image2.