Repeat Images

Dead simple, genuinely useful: this node takes one image and stamps out N copies of it as a single batch. That's the entire job. If you've ever needed to run the same picture through a sampler several times, or feed a batch to a node that only knows how to process batches, this is the two-second fix.

It lives in the art-venture utility drawer alongside the other little quality-of-life nodes. Nothing clever, but the kind of thing you're glad exists when your workflow needs it.

Why you'd reach for it

A few real cases. You want the same input image processed at several different seeds - repeat it to a batch of 4, wire a seed that increments per batch item, done. You're building an img2img or ControlNet pass and a downstream node expects a batch of a certain size - repeat to match. Or you're prepping frames for something that operates over a batch dimension and you need a fixed count of identical starting frames. In all of them the ask is the same: "give me this image, but K times."

It's the counterpart to the batch-splitting nodes: those pull a batch apart, this builds one up.

The inputs and outputs

There are only two inputs and one output, which is the right amount for a node like this:

• images (IMAGE) - the image (or existing batch) to repeat.
• amount (INT, default 1, up to 1024) - how many copies. At the default of 1 the node is a passthrough; set it to 4 and you get four.
• IMAGE (output) - the repeated batch, ready to wire wherever you needed a batch.

If you pass in a batch that already has several images, the repeat applies across it, so mind your final count - it multiplies.

Installing it

ComfyUI Manager, search comfyui-art-venture, install, restart. Or the manual route:

cd ComfyUI/custom_nodes
git clone https://github.com/sipherxyz/comfyui-art-venture

then restart. Pre-installed on comfy.icu.

Common issues

The usual pack-wide one: if this node shows up red in a downloaded workflow, the art-venture pack isn't installed. Manager → "Install Missing Custom Nodes."

The only real footgun is VRAM. A batch is held in memory all at once, so amount: 512 of a large image will happily try to allocate half a gig of tensors and OOM your GPU. Batch size scales your memory use linearly - keep the number sane for your card. If you just want to loop a workflow many times rather than hold every copy in memory simultaneously, a batch isn't the tool; use the queue's batch-count control instead. But for building a modest batch to fan out over seeds or feed a batch-hungry node, this is exactly right.