Random Image Output🧸

Random Image Output takes one incoming image and sends it out one of several outputs, chosen at random. Feed it an image, and on each run it randomly picks which of up to five output branches receives it while the rest stay empty. It's a router with a dice roll baked in.

This is a niche, playful utility. You use it when you want a workflow to fork unpredictably, sending an image through a different downstream path each time you queue.

How it works

You wire an image into the Image input and set randomized_outputs to how many branches you want in play (2 through 5). Each run, the node rolls against the seed and routes your image to exactly one of Image_1 through Image_5, up to the count you chose. The unchosen outputs pass nothing, so whatever's downstream of them simply doesn't fire that run.

The seed behaves like any ComfyUI seed: set it to randomize and you get a fresh pick every queue; fix it and the choice is repeatable, which is handy when you find a routing you liked and want it again.

The inputs

• Image: the single image to route.
• randomized_outputs (2 to 5, default 2): how many output branches to choose among.
• seed: controls the random pick. Randomize for variety, fix to lock a result.

Outputs are Image_1 through Image_5; only one carries the image on any given run.

What it's good for

Think of it as a variety switch for automated or batch runs. Point each output at a different post-processing chain, a different upscaler, a different style pass, and let the workflow surprise you. It's the counterpart to the Image Mixer node in this same pack: Image Mixer shuffles the order of several images across its outputs, while Random Image Output takes one image and gambles on which single branch it goes down. Neither is a workhorse node, they're for adding controlled randomness to a graph.

Installing it

Ships with Plush-for-ComfyUI. Install via ComfyUI Manager (search "Plush", install Plush-for-ComfyUI, restart), or manually: cd ComfyUI/custom_nodes && git clone https://github.com/glibsonoran/Plush-for-ComfyUI.git, then pip install -r requirements.txt, restart.

Common issues

The main thing to understand is that most outputs are empty by design on any given run, so if a downstream branch didn't execute, that's the node working, not failing. If you actually need every branch to run, this is the wrong node. Also double-check randomized_outputs matches how many branches you've wired; setting it higher than the outputs you connected just means the roll sometimes lands on a dead output and nothing happens that run. The pack's How_to_use_RandomizedUtilityNodes.png example workflow shows the intended setup.