Bulin 47 AI Prompt: Copy It for Seedance and Other Video Models

Use a model that accepts a reference video and a reference image in the same request, such as Seedance 2.0 or 2.5. Upload the clip first and the photo second so "reference video 1" and "reference image 1" point to the right files.

Recreate reference video 1 shot for shot as a new vertical 9:16 clip, with the same red-lit restaurant doorway at night, the same composition, the same handheld camera movement and the same timing. The lead performer facing the camera is replaced entirely by the person in reference image 1: that person's face, hair, body and exactly the clothes they wear in reference image 1. Keep the lead's exact gestures, raised hand, head movements and mouth movements from reference video 1, rapping energetically to the beat. Everyone else keeps their position, clothing and movement, but every visible face belongs to a new, different person. Realistic smartphone footage, natural night lighting. No text, no captions, no watermark.