Nodes/ComfyUI-ZML-Image/ZML_预设文本
ComfyUI Node

ZML_预设文本

Five dropdowns of saved prompts, one combined text out

By zml-w·Created about a year ago·Updated 2 months ago· 218
ZML_预设文本
    • 合并文本
    预设1
    预设2
    预设3
    预设4
    预设5

    ZML_预设文本 is a bookmark folder for your prompts. Five dropdowns, each holding a category of saved text, and one combined output that concatenates whatever you selected in each. Select a positive-prompt preset in dropdown 1, a style preset in dropdown 2, a negative in dropdown 3, and the node hands you the merged string to feed into a CLIP Text Encode.

    The source shows it's file-backed, not hardcoded: presets live in a JSON file at zml_w/txt/Preset text/Preset text.json, organized under five fixed category keys (预设1 through 预设5). The dropdown options are rebuilt every time the node loads, so editing the JSON (or using the pack's add-preset endpoints) shows up on the next open. Pick nothing, get nothing - only the categories you select contribute to 合并文本 (combined text).

    The interface

    • 预设1 … 预设5 - five dropdowns, each a category in the JSON file. Default: empty.
    • Output: 合并文本 (STRING).

    That's the whole node. It's a value node in the plumbing sense - one source of text, fanned out, so you're not retyping the same positive prompt in six places.

    Why bother, given the rest of this pack

    The README says the author's own habit is keeping "常用的提示词" - their common Wan prompts - in text-selection nodes, and this is the fixed-category variant. The appeal is curation over speed: five named slots force you to think of your prompt as reusable parts (base prompt / quality / style / negative) instead of one blob you re-type. If that workflow clicks, it's a genuinely nice primitive to chain: feed 合并文本 into the pack's merge or replace nodes and you can assemble prompts programmatically.

    But be honest about the trade. This is the least-flexible of the ZML text nodes - it has no random mode, no per-category free text field, no foldering beyond the five fixed slots. The V3 text-select node (ZML_SelectTextV3) is the one with arbitrary text boxes and folders. Reach for this when five fixed categories is exactly your structure and you want the dropdown discipline; reach for the select nodes when you want power. And the standard reminder: this pack's text widgets are Chinese-first, so the README's translation patch is worth installing for English workflows.

    Install

    Part of ComfyUI-ZML-Image. ComfyUI Manager → search ComfyUI-ZML-Image, or:

    cd ComfyUI/custom_nodes
    git clone https://github.com/zml-w/ComfyUI-ZML-Image
    

    restart. No models, no dependencies beyond the base stack. If the dropdowns ever come up empty after you've edited the JSON, it's almost always a malformed JSON file - the node falls back to an error message rather than crashing, which is helpful, but the console is where you'll find the parse error.

    Categoryimage/ZML_图像/文本

    Inputs (5)

    NameTypeDefaultDescription
    预设1COMBO1 options:
    预设2COMBO1 options:
    预设3COMBO1 options:
    预设4COMBO1 options:
    预设5COMBO1 options:

    Outputs (1)

    NameTypeDescription
    合并文本STRING