🔧 Images Batch Multiple

Core ComfyUI's "Image Batch" node takes exactly two images. If you've got three, four, or five separate image streams you want to stack into one batch, you end up chaining Batch nodes in a little daisy chain - and if the images are different sizes, it just errors. Images Batch Multiple fixes both annoyances: up to five inputs in one node, and it auto-resizes mismatched images so they actually combine.

Why you'd reach for it

Batching is how you push several images through one downstream operation together - one sampler pass, one upscale, one save. The moment you're merging outputs from a few different branches (say four IPAdapter references, or a handful of loaded images) into a single batch, the two-at-a-time core node gets tedious and the size mismatch bites. This node collapses that into a single, readable box, which also keeps large workflows legible - the kind of graph-hygiene win the whole Essentials pack is built around.

How it works

You connect anywhere from one to five images. The node stacks them into a single batch tensor. Because a batch requires every image to share the same dimensions, any input that doesn't match gets resized first - and method is the interpolation used for that resize (lanczos is the sharp default; nearest-exact/bilinear/area/bicubic are the alternatives). It reconciles to a common size and hands you one batch out the other side.

Inputs and outputs

image_1 is required; image_2 through image_5 are optional, so you only wire up as many as you have. method controls the resampling used to make sizes agree - leave it on lanczos unless you're batching pixel art, where nearest-exact avoids blurring. The single output is IMAGE: one batch containing all your inputs, ready to feed a sampler, an upscaler, or a Save Image.

Installing it

Part of ComfyUI Essentials by cubiq (Matteo Spinelli, better known for the IPAdapter node pack and the Latent Vision YouTube channel). Via ComfyUI Manager: search ComfyUI Essentials, install, restart. Or manually:

cd ComfyUI/custom_nodes
git clone https://github.com/cubiq/ComfyUI_essentials

then restart. It appears as 🔧 Images Batch Multiple under essentials/image batch, and the underlying class is ImageBatchMultiple+ - the + is how you tell it apart from the core batch node. The pack's been in maintenance-only mode since April 2025, but this node has no reason to change.

Common issues

Watch the auto-resize. Because it forces every input to one size, batching a portrait and a landscape image together means one of them gets stretched or reinterpreted to match - the batch will run, but the odd one out may look squashed. If that matters, resize your inputs deliberately (Essentials' Image Resize with "pad" or "fill / crop") before this node so you control how the mismatch is resolved. Also remember a batch isn't a grid or a collage - it's N separate images processed together; if you wanted them side-by-side in one picture, that's a different node.