For each skill in muapiapp/skills/library/* (visual, motion, social, edit),
ship a ready-to-load ComfyUI workflow that wires the exact endpoint(s) the
skill uses, so users get a working starting point per skill.
Visual (T2I / I2I): NanoBanana (pro + 2), LogoGenerator, LogoBranding,
YouTubeThumbnail, BlogHeader, InstagramPost, RedNoteCover, KeyboardArt,
AdCreative, BrandKit (3-frame), Storyboard (3 keyframes), AmazonListing,
MultiAngleShots (3 angles), MultiAngleReshoot, PhotoPackGenerator, ActionFigure,
FashionTryOn, UGCLifestyleTryOn, SelfieWithCelebrity, FloorPlanRendering,
InteriorDesign, InteriorVisualizer, CoupleGrid, Brochure, UrlToDesign,
DesignGuide, SocialPack.
Motion (T2V / I2V chains): Seedance2 T2V / OmniReference / FirstLastFrame,
CinemaDirector, DroneStyleVideo, OneShotVideo, 3DLogoAnimation,
AnimalVideoGenerator, CartoonDance, CharacterStoryVideo, GiantProductShowcase,
JewelryProductVideo, MusicVideo (with Suno), ProductAdCinematic,
ProductShowcaseVideo, ProductVideoAdMaker, TalkingBabyVideo, AiFightScene,
UGCAdsWorkflow, SocialMediaVideo, ProductCampaign (multi-stage).
Edit: AiClipping, YoutubeShorts.
3D: ImageTo3D_Tripo, TextTo3D_Meshy (via new MuAPIImageTo3D node).
Generator now also knows MuAPIImageTo3D widget/input/output signatures so it
can be wired into specs the same way as every other node.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
