IO_input_any

You've got a column of seeds, or a list of prompts, or a run of numbers you copied out of a spreadsheet - and you want that as a real list feeding a batch or loop node, not one Primitive node per value. IO_input_any is the pack's answer: paste the text in, tell it how to split it and what type each piece should become, and it fans the result out as a proper list.

How it works

text holds whatever you paste - one block, multiline. split_preset decides how it gets cut up: by line, by space, comma, period, semicolon, tab, or pipe, or Custom if you want to hand it your own separator in the delimiter field. None means it doesn't split at all. Once it's split, output_type decides what each piece gets coerced into - int, float, string, boolean, dictionary, set, tuple, or the default anytype, which just leaves values untyped so the wire can go into whatever expects them.

Inputs and outputs

• text - the raw text block.
• split_preset - how to cut it (Line/Space/Comma/Period/Semicolon/Tab/Pipe/Custom/None).
• delimiter - your own separator, used only when split_preset is Custom.
• output_type - what type to cast each item to.

Two outputs: data, a list (each split, typed item fanned out - its type is a wildcard * so it'll connect to whatever slot is waiting), and list, the same data bundled as a single LIST object rather than a fan-out. Use data when the next node wants one value per batch item; use list when the next node wants the whole collection handed over at once.

How to install it

ComfyUI Manager → search ComfyUI-Apt_Preset, install, restart. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/cardenluo/ComfyUI-Apt_Preset.git

then run install.bat for the pack's dependencies (it's Windows-only - on Linux/macOS, open it and run the pip install lines by hand) and restart.

Common issues & troubleshooting

Picking dictionary, set, or tuple on plain lines of text. Those types expect the text to actually be shaped like the Python literal - a dictionary needs key/value structure, not just a list of words. If you get a parsing error on output_type, that's almost always the cause; drop back to string, int, float, or anytype unless you know your text is genuinely formatted for one of the structured types.

Nothing splits when you expected it to. Check split_preset isn't sitting on None (which does nothing) and, if it's Custom, that delimiter actually has something in it - an empty custom delimiter with Custom selected won't fall back to a sensible default.

IMPORT FAILED on load. This pack has at least one reported import failure in the wild - usually traced back to skipping install.bat's dependency step, especially off Windows where you have to run its pip install lines manually. Check the console log for the actual missing module before reinstalling.