Showcases

Storyboard Concept Film: From a Single Description to First and Last Frame Images

Using seedance-2.0-mini's text-to-video pipeline for a "storyboard concept" demo: first generate a first-frame and a last-frame reference image from two visual descriptions, then bind the two images as first_frame_url and last_frame_url, letting the model complete a continuous camera move from a morning platform to a departing train within a 16:9, 8-second interval. First-and-last-frame constraints pin down the hardest thing to express in words—the opening and closing framing—while the middle is left to the model to fill in. This suits ad previsualization, storyboard validation for narrative shorts, and quickly showing non-technical colleagues the path from text to image. Keep the prompt within a single shot to avoid subject drift from multi-scene switching; enable native audio to get an ambient sound bed synced to the picture.

Style Storyboard · Cinematic · Text-to-Video

Model Seedance 2.0

Reference inputs

last_frame_url
first_frame_url

Try it yourself

Prompt

Try this model

A single continuous cinematic long take, 16:9. Opening image: at dawn, a lone commuter stands at the far end of an empty urban train platform, a long shadow slanting across the wet concrete floor, cold blue ambient light interwoven with warm sodium lamps, shallow depth of field, 35mm anamorphic widescreen look. The camera then slowly trucks left at eye level, easing in slightly as a train arrives and glides to a stop; steam and dust drift through the light beams, reflections stretch across the platform floor. Closing image: the doors open, the commuter has stepped inside, warm carriage light spills onto the platform. Teal-and-orange color grading, fine film grain, subtle lens flare, steady motion, no cuts, natural pacing.

Copy the full prompt and adapt it for your own generation.