Nodes/Sage Utils/Pony Source
ComfyUI Node

Pony Source

Source_anime or source_furry, chosen from a dropdown

By arcum42·Created 2 years ago·Updated 28 days ago· 33
Pony Source
    • source_text
    sourcenone

    Pony is one model that does a lot of styles, and the source tags are how you tell it which one you mean. source_anime for anime, source_furry for the furry art, source_cartoon for western cartoons, source_pony for the actual My Little Pony look - plus a few more. It's Pony's style dial, and it sits right after the score tags in the standard prompt header.

    Pony Source turns that dial into a dropdown. Pick a source, get the exact source_x, fragment, comma and space included, ready to join into your prompt.

    How it works

    One combo input, source, with the full list: none, pony, furry, anime, cartoon, 3d, western, comic, monster. Choose none and the output source_text is empty - handy when you're building a prompt that shouldn't carry a source tag at all. Choose anything else and you get source_{source}, .

    That trailing comma and space aren't cosmetic. Pony's special-tag header reads as one continuous string - the score tags, source tag and rating tag typically sit together at the front of the positive prompt - and the pack's Pony nodes all emit fragments with the same punctuation so they chain cleanly through a Join Text node.

    Why the source tag earns its place

    The source tag is genuinely load-bearing. The KB's Pony notes quote the community's own advice - put source_anime in the positive prompt and push source_furry, source_cartoons into the negative to keep a generation from drifting styles. Because Pony's training spans so many art genres, leaving the source tag out is how you get a prompt that returns "4 different styles" across seeds. A dropdown doesn't make that insight for you, but it makes acting on it painless.

    Wiring it in

    source_text is a STRING. Combine it with Sage_PonyScore and Sage_PonyRatingv6 (or the all-in-one Pony Prefix node) to build the full Pony header before the prompt hits a CLIP text encode. The natural pairing with the pack's Join Text / Triple Join Text nodes is right there.

    Installation

    Sage Utils install, same as the rest of the pack. ComfyUI Manager, search "Sage Utils", or:

    cd ComfyUI/custom_nodes
    git clone https://github.com/arcum42/ComfyUI_SageUtils
    cd ComfyUI_SageUtils
    pip install -r requirements.txt
    

    Restart ComfyUI. Only dependency is dynamicprompts.

    The honest take

    Like the other Pony helpers in this pack, it's a string formatter with a vocabulary list behind it - you could type source_anime, yourself. But the vocabulary is easy to misremember (is it source_cartoon or source_cartoons?), and when a tag is misspelled it does nothing at all. The dropdown removes that failure mode and keeps your Pony headers consistent across every workflow. If you're already running Sage Utils for its metadata or LoRA nodes, this is a free upgrade to your Pony prompt hygiene.

    CategorySage Utils/text/prompt/pony

    Inputs (1)

    NameTypeDefaultDescription
    sourceCOMBOnoneChoose the prompt source type for Pony prompts.

    Outputs (1)

    NameTypeDescription
    source_textSTRINGThe source text fragment generated for Pony prompts.