imageSplitList

Wire in one image batch, get three image outputs back. That's the shape of this node: a single images input and three images output slots, no other controls exposed on the node itself. If you've got a batch and you want it broken across three parallel downstream paths - three different post-processing chains, three separate saves, three branches of a comparison - this is what does the splitting.

Worth knowing up front: the pack's own changelog describes this node's original purpose as splitting a batch "every N images," which implies an adjustable group size. The current node, though, exposes no count or group-size control at all - just the one input and three fixed outputs. Treat this as "splits your batch across three output slots" rather than assuming you can dial in an arbitrary N; if you need a specific, adjustable split count, this isn't the node that gives you a widget for it.

Why you'd reach for it

The practical use is fan-out: you have one batch and three things you want to do to it in parallel, without gluing them all onto the same downstream chain. Feed the batch in once, get three handles to work with out, and run each through its own path - maybe one gets upscaled, one gets saved as-is, one gets a detail pass. It saves you from having to duplicate the batch manually before branching.

The inputs and outputs that matter

• images - the batch to split. The only input.
• Three images outputs - the batch distributed across three slots. They're unlabeled beyond their position, so treat them as "output 1, 2, 3" and connect whichever ones you actually need; you don't have to use all three.

Installing it

Ships with the base pack. ComfyUI Manager: search ComfyUI Easy Use, install, restart. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/yolain/ComfyUI-Easy-Use

then install.bat on Windows or pip install -r requirements.txt, restart. No models needed.

Common issues & troubleshooting

Not sure which images landed on which output. Since the three outputs carry no labels beyond their slot position, and there's no visible control for how the split is decided, the safest habit is to test with a small batch first - feed in three or four clearly distinguishable images and preview each of the three outputs individually to see how they're distributed before you trust this node in a larger graph.

Expected an adjustable group size and don't see one. As noted above, the node's current inputs don't expose a count parameter, even though the pack's history describes it as an "every N images" splitter. If you need precise, adjustable control over how a batch gets divided, easy imageSplitGrid (row x column, for a single image) or easy imagesSplitImage (up to five explicitly named single-image outputs from a batch) are the nearby nodes worth checking instead - one of them is more likely to match what you're actually trying to do.

Leaving an output unconnected. That's fine - you don't need to wire all three. An unconnected output slot just doesn't run that branch; it doesn't error and it doesn't cost you anything.