Image Batch To List ♾️Mixlab

ComfyUI treats a batch and a list of images very differently under the hood: a batch runs everything downstream once, on the whole stack at once, while a list re-executes the downstream chain once per item. Most of the time a batch is what you want - faster, one pass. But sometimes you need per-image behavior: saving each with its own filename, running a node that only makes sense one image at a time, or feeding something that just doesn't handle batches well. Image Batch To List is the standard "flip the switch" node for that - Mixlab's own implementation of a pattern most packs ship some version of.

How it works

It's about as simple as this pack gets: take a batched IMAGE tensor in, and split it into a Python list of individual images out. Everything wired after this node now re-runs once per image in that list, rather than once for the whole batch.

The inputs and outputs that matter

• image_batch (IMAGE, required) - the batched images to split.

The single output is image_list (IMAGE, marked as a list) - wire it into whatever needs per-item execution downstream.

Its companion in this same pack, ImageListToBatch_, does the reverse - collapsing a list back into a single batch once you're done processing items individually.

How to install it

Ships with the Mixlab pack.

• ComfyUI Manager: search comfyui-mixlab-nodes, install, restart.
• Manual: cd ComfyUI/custom_nodes && git clone https://github.com/shadowcz007/comfyui-mixlab-nodes.git, install requirements (install.bat on Windows, or ../../../python_embeded/python.exe -s -m pip install -r requirements.txt), restart.

No model downloads - this is a pure batch/list conversion utility.

Common issues & troubleshooting

Queue suddenly takes much longer. That's expected, not a bug - converting to a list means everything downstream re-runs per image rather than once for the whole batch. Fine for a handful of frames; noticeably slower (and more GPU-expensive, if downstream nodes are heavy) for a large batch. Only convert when you actually need per-item behavior.

I only wanted to peek at individual images. If you're not actually doing per-item processing, you probably don't need this conversion at all - a batch already previews as individual thumbnails in most viewer nodes.

Need to recombine afterward. Route the list straight into ImageListToBatch_ once your per-item processing is done, to get back to a single batched tensor for anything downstream that expects one.

This is one of those nodes you won't think about until you need it, and then you'll need it constantly - anywhere in this pack's Layer/Grid tooling where a per-region or per-frame operation has to happen individually before recombining, a batch-to-list step like this one is usually sitting quietly in the middle of the graph.