Superside Prompt Splitter

Somewhere in the Superside pack there's a node that takes up to ten reference images and one prompt and does something with them. The question nobody wants to answer by hand is: how do you feed ten different prompts into those ten slots without building ten separate text boxes? Superside Prompt Splitter is the answer - you type red dress * blue dress * black dress into one box and get three separate STRING outputs.

How it works

It's a string operation with a UI. You give it a single multiline prompt and a separator symbol - the default is * - and it splits the text on that symbol, trimming whitespace, and hands each piece out as its own output: text_1 through text_10. Unused slots come back as empty strings, which is deliberate: a downstream node with an empty prompt does nothing, so you can leave a slot dangling without breaking anything.

The practical thing to remember is that the separator is a literal character, not a regex. If your prompt legitimately contains an asterisk - say a mathy product description - change the separator to something you won't accidentally type, like | or ###. The tooltip spells this out too, which is the author's way of telling you the * default is a suggestion, not a law.

After each run the node displays the split result in its UI as a numbered list, which makes it easy to see at a glance whether the separator actually caught where you thought it did.

Inputs and outputs that matter

• prompt - the multiline text to split.
• separator - the split symbol, default *. This is the one you'll actually touch.
• Outputs text_1…text_10 - one STRING per split segment, empty strings for unused slots.

There's no model, no API call, and no api_key input - a pure local utility, like most of this pack's text nodes.

Installing it

It ships in comfyui-superside-nodes, so install the pack once:

cd ComfyUI/custom_nodes
git clone https://github.com/Superside/comfyui-superside-nodes.git
cd comfyui-superside-nodes
pip install -r requirements.txt

Restart ComfyUI and it appears under the Superside category. No extra dependencies beyond what the pack installs (fal-client, pillow, numpy, torch, requests), no model files to download.

Where you'd actually use it

The intended pattern, straight from the node's own description, is feeding individual prompts into multi-image nodes - the pack's Seedream V4.5 Edit and Seedream V5 Pro Edit take up to ten reference images, and the Seedance Lite image-to-video node takes up to four. Write four short scene descriptions separated by *, split them, and each lands in its own image_N-adjacent prompt slot. It's also handy any time you want to compose a batch of related text fragments - captions, negative-prompt lists, style snippets - and fan them out to different consumers without a dozen hand-placed text nodes.

The one thing it won't do is split into an existing list-type input. If the node you're feeding expects a single STRING list rather than ten individual STRING ports, you'll need to join them yourself. For the common slot-per-output case, though, it's exactly the tool. Keep the separator unique, keep an eye on the numbered UI preview, and it'll behave.