Generating six consistent characters in one image: the practical limits of Illustrious and Anima
Jul1an_Gut1errez_777 · reddit · 2026-09-19
A user working with Illustrious and Anima wants to generate six different characters in a single scene with distinct poses, per-character consistency, and high quality. They run into the usual problems — attribute bleeding, degrading faces, characters merging — and regional prompting hasn't helped. They ask whether it's feasible, what the best approach is, whether one LoRA per character can work without interference, and which ComfyUI/Forge extensions and workflows to use.
More from Multimodal
- Qwen Image 2.1 PR lands in ComfyUI, open weights look nearly certain — Altruistic_Heat_9531 · 2026-09-19
- Director's Console update adds model-specific prompt guides, goes MIT licensed — npittas · 2026-09-19
- Magnific founder's secret 6-month AI-generated series 'Backrooms Nemoris' premieres — Loo_Atreides · 2026-09-19
- Prompting AI video to look real: ask for a human's phone-camera mistakes, not perfection — socialwithaayan · 2026-09-19
- Flux 2 Klein struggles with style+palette transfer from reference images — wacomlover · 2026-09-19
- Dev spends two weeks having Astra 3D-model every object in a three.js scene — danbri · 2026-09-19