Showcases

From Storyboard Concept to Frame: Text-to-Video Reference Generation

Using "from text to image" as its premise, this piece demonstrates how a reference-to-video pipeline uses multiple reference images to constrain subject, scene, and style, then turns a storyboard concept into a coherent video. The work first uses two reference assets to lock down the character's appearance and the environment's color tone (one character design, one scene atmosphere); prompts then describe camera staging and plot progression, outputting an 8-second 16:9 final cut. It suits early-stage workflows that need to "land a text script as visuals first," such as storyboard previz, script visualization pitches, and ad concept validation. Reference images carry consistency and art direction, while prompts carry narrative and camera movement; the division of labor is clear, making it easy for developers to reuse as a template.

Style Storyboard · Concept Art · Narrative

Model Gemini Omni

Reference inputs

image_urls
image_urls

Try it yourself

Prompt

Try this model

Bring a storyboard concept to life as a short cinematic clip. A young explorer in a distressed olive field jacket stands at the edge of a rain-soaked stone courtyard, looking up at a towering brass observatory dome. The camera starts low and close on the explorer's boots in a shallow puddle, then cranes upward in one smooth continuous rise to reveal the dome's silhouette against cool gray morning light. The camera pushes in slowly as the explorer steps forward and raises one hand toward the structure; raindrops are rim-lit and mist drifts across the flagstones. The palette is muted teal and warm amber, with natural volumetric fog, soft overcast key light, and a warm practical glow from the dome's windows. The pacing is steady and unhurried; a subtle handheld sway gradually transitions into a stabilized crane move. Fabric folds, water splashes, and drifting mist all carry physically accurate realism.

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