Character Swap inverts the usual prompt-first image workflow: instead of describing desired lighting, mood, and composition in text, find an existing photo that already has that look and swap a trained identity into it. The user supplies a Soul ID: Personal Likeness Training from Reference Photos-trained character as the source and a separate target photo — one that already contains a person in the desired scene — and the model replaces that person while preserving the target's lighting, mood, composition, and color tones.
This turns scene-matching into a search problem rather than a prompt-engineering problem: shoot or source reference photos with the right vibe first, then let Character Swap carry the trained identity into them. Demonstrated as one stage of a four-step pipeline built on Soul 2: Production-Ready Start-Frame Image Model (Soul 2.0) — base generation, Soul ID training, Character Swap, then prop inpainting — for building a consistent AI-influencer / UGC-ad content workflow.