Nodes/Sage Utils/Pony Rating (v7)
ComfyUI Node

Pony Rating (v7)

Pony v7's rating tags, without you mixing up safe and general

By arcum42·Created 2 years ago·Updated 28 days ago· 33
Pony Rating (v7)
    • rating_text
    ratingnone

    Pony v7 changed the rating vocabulary, and that's exactly the kind of detail that costs you an evening. V6 used rating_safe, rating_questionable, rating_explicit. V7 - the AuraFlow-based sequel that shipped in late 2025 - uses rating_general, rating_sensitive, and rating_explicit. "Safe" became "general", "questionable" became "sensitive", and if you're copying old V6 prompts into a V7 workflow, you're pasting a vocabulary the model half-recognizes.

    Pony Rating (v7) is the v7-flavored version of the pack's rating generator: pick the level from a dropdown, get the exact tag string the author mapped for V7, comma and trailing space included.

    How it works

    One combo input, rating, with the V7 list baked into the node: none, general, sensitive, explicit. none outputs an empty string; anything else outputs rating_{rating}, .

    That's the whole mechanism, and it's the right size. The hard part was the vocabulary, not the string building, and the node just prevents the mistake. The author ships two rating nodes precisely because the two Pony generations don't share a rating scheme - the tooltips even call out "(v6 style)" and "(v7 style)" so you grab the right one.

    Why the vocab matters

    The rating tag is part of Pony's special-tag header, the same family as the score tags and source tags. The KB's Pony notes cover the mechanics: these tags control the SFW/NSFW dial, and Pony's training made the exact tag strings matter more than they should. On V7 the model card itself says the quality tags weakened, but the rating system is still how you steer content level.

    A real-world gotcha the community has hit: a lot of V7 users still reach for rating_safe out of habit, and some V7 generations were shared with the old V6 vocabulary embedded - it "works" because the model fuzzily understands, but you're betting on a fuzzy match instead of the tag the author actually mapped. This node removes that bet.

    Installation

    Same pack as the whole Sage Utils suite. 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

    If you only ever generate Pony V6 - and honestly, that's most of the Pony world, since V7 never built momentum - you don't need this node; grab the v6 sibling instead. But the moment you're testing V7 and its new prompt format, this is the piece that keeps your rating tags on the current vocabulary instead of the one you have memorized. It's a three-word prompt fragment, yes. It's also the difference between "why is this V7 output drifting" and a prompt header you can stop thinking about.

    CategorySage Utils/text/prompt/pony

    Inputs (1)

    NameTypeDefaultDescription
    ratingCOMBOnoneThe Pony v7 rating level to include.

    Outputs (1)

    NameTypeDescription
    rating_textSTRINGThe generated Pony v7 rating text.