Concept art to rigged character: skill partitioning + ComfyUI + Trellis pipeline demoed
majidmanzarpour · x · 2026-10-09
A developer working in @spawn demoed a full pipeline turning concept art into a rigged, animatable game character:
- First, a skill by @majidmanzarpour partitions the concept art into pieces (body and clothing separated)
- The pieces go through ComfyUI + Trellis to produce a skinned mesh: the body gets auto-rigged while clothing gets spring bones
- Body and clothing remain independent objects, with clothing parented to the model; invisible collider balls prevent clipping
The author notes it's not AAA quality yet and the work is ongoing, but the pipeline is a solid reference for AI-assisted 3D character generation.
More from Multimodal
- Seedance 2.5 AI video praised as a "masterpiece" in viral share — BLUECOW009 · 2026-10-09
- HeyGen Video Tops OpenRouter's Video Rankings, Hands Out Codes for 100 Free Clips Each — aziz4ai · 2026-10-09
- Building 'Night Signal', a lost-media horror broadcast, with Claude, Codex and a custom Canvas pipeline — zndr-cs · 2026-10-09
- One Prompt, All Code: Claude Motion + Higgsfield Katana Demo Wows Users — SimplyAnnisa · 2026-10-09
- User One-Shots a Higgsfield Supercomputer Launch Video with Higgsfield Katana — SimplyAnnisa · 2026-10-09
- An 1890s photo of Cecil Rhodes brought to life with AI wins first prize — aeeiee · 2026-10-09