Exploring Workflows for Consistent AI Character Identity and Body Swap
GooDroop · reddit · 2026-08-05
A creator is seeking the optimal workflow for achieving consistent AI character identity (face + body) for realistic UGC content in 2026. The goal is to map a custom AI character onto reference images with varying poses, clothing, and compositions. Having faced setup issues with PuLID + Flux, the author asks the community whether newer models like Krea 2, Flux.2, and InfiniteYou, or training a custom LoRA, are better suited for a production environment.
More from Multimodal
- Fable Multimodal Model Generates Delightful Short Film: What the Table Does at Night — repligate · 2026-08-05
- Fable Model Generates Stream of Consciousness Artwork: Condensation — repligate · 2026-08-05
- MiniMax H3 Test: Achieves Accurate Voice Cloning via Audio Reference Workflow — Perfect-Campaign9551 · 2026-08-05
- Fun AI Video: Generating a Crossover Scene of Friends and The Office via Prompt — Ciko15 · 2026-08-05
- MONET Dataset Released: 105M Samples for Open Text-to-Image Research — victormustar · 2026-08-05
- 95-Minute AI Feature Film Made for $500K Fully Open-Sourced — CurieuxExplorer · 2026-08-05