Step 1: Selecting the Ideal Character Image
Your source image serves as the identity anchor. For best results, use an evenly lit photo with neutral contrast where the character's facial features, hands, and outfit are clearly distinguishable. Front-facing or three-quarter poses from head to waist (or full body) yield the cleanest motion continuity.
Step 2: Choosing Reference Choreography
The motion reference dictates the timing, footwork, and energy of the synthesized video. Select from our built-in viral choreography library (Hip-Hop Groove, K-Pop Viral Chorus) or upload your own clean reference video. Avoid source videos with extreme camera shake or multiple overlapping dancers.
Step 3: Genjutsu AI Prompt Formulas for Video-to-Video
When guiding motion transfer models, prompt specificity matters. A proven prompt template structure is: [Subject Description] + [Dance Style & Tempo] + [Lighting & Environment] + [Camera Lens Details]. Combining clear reference imagery with targeted motion prompts produces maximum consistency.
Step 4: Exporting for TikTok, Reels, and YouTube Shorts
Genjutsu AI Generator renders directly into vertical 9:16 HD MP4 at 1080x1920 with synchronized audio tracks, eliminating post-production cropping. Download the MP4 file and publish directly to mobile video platforms.