ComfyUI Node

String OutputList

Turn a text block into a list of prompts (and steal any dropdown's options)

By geroldmeisinger·Created 9 months ago·Updated 2 days ago· 184
String OutputList
    • value
    • index
    • count
    • inspect_combo
    separator\n
    values

    This is the workhorse of the OutputLists Combiner pack, and honestly the node you'll reach for most. You paste a block of text - one prompt per line - and String OutputList splits it into an output list, so every node downstream runs once per line. Four prompts in the box, four images out of the KSampler, zero fiddling. It's the closest thing to Automatic1111's prompt batches in ComfyUI, and a common verdict on the author's launch thread was that it "feels like using Automatic1111 again."

    How it works

    Two inputs, both simple:

    • separator - what to split on. Defaults to \n (the string is unicode-unescaped, so \n really means newline). You can split on , or ; or ### if that's your thing.
    • values - the multiline text box with your prompts.

    It trims trailing newlines before splitting and strips whitespace around each item, so stray blank lines don't produce empty list entries. Outputs: value (the output list - wire this to a prompt), index (0..count, handy for labels), and count (a plain integer).

    The input that makes it special: inspect_combo

    There's a fourth output, inspect_combo, and it's the feature the author says he couldn't find anywhere else. It's a dummy COMBO output. Connect it to any dropdown - the LoRA loader's lora_name, the KSampler's sampler_name, scheduler - and it pre-fills the values text box with that dropdown's full option list. The connection then auto-relinks to value. You get "iterate every sampler" as a one-drag operation: link inspect_combo to the sampler dropdown, trim the list to the ones you care about, and now the sampler sweep runs itself. No typing model names by hand, no custom sampler nodes.

    Installing

    Search "OutputLists Combiner" in ComfyUI Manager, or:

    cd ComfyUI/custom_nodes
    git clone https://github.com/geroldmeisinger/ComfyUI-outputlists-combiner
    cd ComfyUI-outputlists-combiner
    uv pip install -r requirements.txt
    

    Then restart. The pack's heavy deps (pandas, skia-python) install once and are shared by all its nodes.

    Common issues

    The "I only got one image out" confusion hits here hardest, because nothing about the UI screams "this runs multiple times." Each line produces its own image with batch_size=1; if you want them collected, add Rebatch Images before your grid node.

    For debugging, the pack README recommends pythongosssss's Show Text node: most string-display nodes only show the first or last list item, but Show Text prints a new entry for every emitted item. Cheap sanity check before you burn a 500-run queue on a typo in line 3.

    One gotcha if you change separators: values are whitespace-trimmed per item, but a separator like , inside a prompt (someone always has one) will silently split mid-prompt. When in doubt, keep the newline default.

    CategoryUtility

    Inputs (2)

    NameTypeDefaultDescription
    separatorSTRING\nThe string used to split the textfield values by.
    valuesSTRINGThe text you want to split into a list. Note that the string is trimmed of trailing newlines before splitting, and each item is again trimmed of whitespace.

    Outputs (4)

    NameTypeDescription
    value*The values from the list. use(s) `is_output_list=True` (indicated by the symbol `𝌠`) and will be processed sequentially by corresponding nodes.
    indexINTRange of 0..count. You can use this as an index. use(s) `is_output_list=True` (indicated by the symbol `𝌠`) and will be processed sequentially by corresponding nodes.
    countINTThe number of items in the list. use(s) `is_output_list=True` (indicated by the symbol `𝌠`) and will be processed sequentially by corresponding nodes.
    inspect_comboCOMBOA dummy-output you can use to link to a `COMBO` and pre-fill with it's values. The connection will then be automatically re-linked to `value` output.