Create a Continuous Street-Dance Video from a 12-Panel Storyboard

Create a dance video using the following specifications:
- Character: ${character}
- Outfit: ${outfit}
- Location and visual style: ${location-style}
- Dance style: ${dance-type:hip-hop street dance}
- Total duration: ${duration:30 seconds}

First, use ChatGPT to choreograph the complete dance and divide it into 12 storyboard panels. Keep the character, outfit, and location consistent throughout. Each panel should clearly specify only these storyboard essentials: camera framing or position, the dancer’s body pose, and camera angle or movement. Make the choreography precise, including footwork details such as whether the heel or ball of the foot lands first.

For every transition between panels, describe how the movement flows from one to the next (for example: step down → shift weight → explosive move → return to position → enter the next move). The ending pose of each panel must match the starting pose of the next. Give each shot a clear purpose: use a wide shot to show full-body choreography, a low angle for footwork, a close-up for a specific movement, or a tracking shot to follow the dancer’s path.

Create the 12-panel storyboard, then use Seedance 2.5 to generate the corresponding video shots in order, following the storyboard and preserving the specified character, outfit, location, choreography, camera direction, and lighting. If a previous clip is supplied for extending the video, use it as the continuity reference: match its ending pose, camera movement, lighting, and direction before continuing the choreography.

Avoid teleporting, arbitrary spins, sliding steps, body deformation, or abrupt camera changes.