🔧 Image List To Batch

ComfyUI has two ways of carrying more than one image down a wire, and they are not interchangeable. A list is N separate images that nodes process one at a time. A batch is a single tensor with all N images stacked inside it. Plenty of nodes hand you a list - loaders that grab a folder, splitters, anything with a list output - and then the node you want to feed next expects a batch, so it either errors or quietly only touches the first frame. ImageListToBatch is the adapter between those two worlds.

It comes from ComfyUI Essentials by cubiq (Matteo Spinelli - the same dev behind the reference ComfyUI_IPAdapter_plus pack), a collection he describes as the nodes "weirdly missing from ComfyUI core." This is one of the missing plumbing pieces: obvious once you need it, annoying that you have to install a pack to get it.

How it works

A batch tensor has one hard rule - every image in it must be the same width and height, because they're literally stacked along a new dimension. A list has no such rule; the images can be all different sizes. So the node can't just concatenate blindly. When your list has mismatched dimensions, it resizes them to a common size before stacking, then returns the whole thing as a single batched IMAGE. If your images are already uniform (say they all came out of the same generation), it just stacks them and you're done.

The inputs and outputs that matter

There's exactly one input and one output, which is the whole appeal:

• image (IMAGE) - the incoming list. Wire your list-producing node straight in.
• IMAGE (output) - a batch. This goes into anything that wants all the frames together: Save Image (to write them as a set), a video-combine node, a VAE encode you want to run in one pass, or a preview that shows the whole strip.

No settings to tune. It either fits your graph or it doesn't.

How to install it

Easiest path is ComfyUI Manager: open it, search ComfyUI Essentials, install, restart. Manual clone works too:

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

then restart ComfyUI. No models to download for this one - it's pure tensor reshaping.

Common issues & troubleshooting

It resized my images and I didn't want that. That's the node doing its job - a batch physically can't hold mixed sizes. If you need the originals preserved at their own dimensions, you want a list, not a batch, so rethink whether the downstream node actually needs batching or whether you should keep it a list and let it iterate.

The node isn't there after a ComfyUI update. Essentials has been in maintenance-only mode since April 2025 - cubiq stepped back from ComfyUI, and the community has watched individual nodes in the pack break after big Comfy frontend updates. If it vanishes or throws on load, update the pack through Manager first; if that doesn't fix it, you're looking at version drift, and rolling ComfyUI back to a known-good build is the usual workaround people land on.

You actually wanted the opposite. If a node downstream complains that it got a batch when it wanted to process images individually, reach for ImageBatchToList - the mirror node in the same pack.