Nodes/ComfyUI Text-Driven Workflows/Single Prompt (Text-Driven)
ComfyUI Node

Single Prompt (Text-Driven)

One prompt fragment, one weight, one switch — the pack's building block

By kanryu·Created 4 months ago·Updated 3 months ago· 0
Single Prompt (Text-Driven)
    • string_out
    text
    multiplier1.0
    activetrue
    optional_text_in

    Single Prompt (Text-Driven) is the simplest node in this pack, and it's honest about it: one text box, one multiplier, one on/off switch, one string out. It's the per-fragment control you drop in when you don't need Multi Prompts' ten slots or Text Line Prompt's line editor - you just want one prompt piece you can weight and flip without editing text by hand.

    How it works

    Three inputs do the work:

    • text - your prompt fragment, multiline.
    • multiplier - a float from 0.0 to 10.0, default 1.0. Anything other than 1.0 wraps the fragment in CLIP attention syntax, so 1.5 becomes (your text:1.5).
    • active - the master switch. Flick it off and the node outputs an empty string instantly, no deleting required.

    There's also optional_text_in, an optional string pin. If you wire something into it, it gets appended to text with proper spacing - this is how you build a fragment from multiple sources without gluing strings manually. The output is string_out, ready to feed a Join Strings node or go straight into CLIP Text Encode.

    It's a thin node and it knows it. But thin has a job here: the pack's whole design is "decouple the character core from the staging direction," and this is the unit you compose those from. Several Single Prompt nodes feeding a Join Strings gives you per-fragment weights and per-fragment muting, which is a surprising amount of control for zero graph clutter.

    Where the weight thing bites

    The multiplier wraps text in (prompt:weight) syntax, which is only meaningful on CLIP-encoded models - SD 1.5, SDXL, Illustrious, Pony and friends. On Flux, and especially the LLM-encoded models (Z-Image, Klein, Anima, Krea 2), attention weights are discarded by the encoder, so your 1.5 does nothing but add parentheses to an instruction. Use the weight for what it is: a CLIP-era tool.

    Install

    Comes with ComfyUI Text-Driven Workflows. ComfyUI Manager → search "ComfyUI Text-Driven Workflows" → Install, or:

    cd ComfyUI/custom_nodes
    git clone https://github.com/kanryu/ComfyUI-Text-Driven-Workflows.git
    

    Restart. No pip dependencies, no model files. The pack does require a current ComfyUI with the V3 extension API - on an older install the whole pack fails to load (check the console for "Failed to load nodes") until you update.

    Bottom line

    If you're already using the pack's other nodes, this slots in as the clean single-purpose fragment holder. On its own it's a fancy text passthrough with a weight and a mute button - but that's precisely the niche it fills, and it fills it without ceremony.

    CategoryText-Driven Workflows

    Inputs (4)

    NameTypeDefaultDescription
    textSTRING
    multiplierFLOAT1.00–10
    activeBOOLEANtrue
    optional_text_inoptSTRING

    Outputs (1)

    NameTypeDescription
    string_outSTRING