Showcase

Example
12.0slength2scenes

The brief

  • Format: Story.
  • Visual style: Photorealistic cinematic.
  • Framing: 9:16 — Shorts, Reels, TikTok.
  • Target length: 24s.
  • Narration language: English (US).
  • Subtitles: burned in.

Published

The cast

Each one was generated once as a reference image, then reused in every shot it appears in.

  • falling_woman — Create a single photorealistic cinematic identity reference portrait/full-body character sheet of a young adult woman with long medium-brown hair, warm natural complexion, expressive brown eyes, and a believable everyday face. She wears a plain tan long-sleeve shirt, fitted black pants, and simple dark shoes. Show her standing neutrally in a clean uncluttered studio setting with soft even daylight, hair visible, full outfit clearly readable, realistic proportions, natural skin texture, restrained makeup, and no props, logos, text, or background location. This is an identity anchor for later freefall and storefront video; preserve the same face, hair length and color, tan shirt, and black pants. Appears in 2 shots.

Shot by shot

The finished film runs 0:24 across 2 shots.

  • Shot 1 — Image 1 is the sole human identity reference for the clip: preserve this exact young woman@Image 1, with long brown hair, tan shirt as the dominant and unmistakable upper garment in every frame, black pants, natural face, and consistent proportions. Image 2 is only an aircraft/environment reference: use its large cargo airplane, open rear ramp, cloud deck, blue daylight, and altitude; ignore and replace every human figure visible in Image 2. The only human in the entire clip is the woman from Image 1, and she is never dressed in black. Shot 1: Medium-wide view from behind and slightly above the open cargo ramp, close enough that the woman's tan shirt is clearly visible as she falls below the cargo airplane into the cloud deck, arms stretched outward, long brown hair streaming backward, eyes wide and mouth open, terrified but unharmed. One smooth camera pull-away establishes altitude. The aircraft structure from Image 2 stays overhead; use no person from Image 2. <powerful continuous wind, distant aircraft engines> Shot 2: Medium close front-facing view of the same woman@Image 1 suspended in freefall against bright clouds and blue atmosphere. Her tan shirt fills the torso area and remains clearly visible. The camera holds steady relative to her face. She keeps her arms extended, breathes against the wind, looks directly into the lens, and says in a natural young-adult female American English voice, clear, conversational, slightly breathless under strong wind {Most people think you’d want to land in water. That’s the worst option. Water doesn’t compress, and you’d hit it at over 200 miles per hour. And that’s the same as landing on concrete.} Shot 3: First-person view from the same falling woman@Image 1, her hands and black-pants legs visible at the edges as the ocean expands far below through broken clouds. A smooth downward tilt reveals the water surface; keep the same blue daylight and realistic atmospheric scale. <continuous rushing air> Shot 4: The first-person camera from the same woman@Image 1 rotates from the ocean toward a distant city grid of rooftops, roads, and neighborhoods below, keeping one continuous horizon and blue daylight. End facing the urban direction that will lead into the next scene. Avoid any other humans, black tops on the woman, parachutes, injuries, invented landmarks, and any generated text or subtitles inside the footage. Photorealistic cinematic imagery, realistic human motion and wind, crisp natural daylight, subtle film grain, vertical social-video composition, no music. Features falling_woman.
  • Shot 2 — Image 1 is the identity reference for the same falling woman: preserve her exact face, long brown hair, tan shirt, black pants, and proportions. Image 2 is the supplied Cabinet Guys LLC storefront reference in Knoxville, Tennessee: match its brick facade, green metal roof, white sign band, blue-and-red Cabinet Guys LLC sign, entrance, columns, shrubs, parking lot, and daylight; treat Image 2 as the exact destination and preserve the business sign visibly and legibly. Image 3 is a supplied location-and-wording reference showing the Knoxville location context and the wording Custom Cabinets Knoxville, TN — Knoxville, Tennessee; use it only to confirm the spoken location, never as a visual overlay or replacement storefront, and do not reproduce its search UI. Shot 1: First-person view from the same falling woman@Image 1, hands and feet briefly visible as rooftops, roads, and neighborhoods pass below. The camera makes one controlled accelerating zoom downward and locks onto the exact Cabinet Guys LLC storefront@Image 2; use only the supplied destination, with no substitute buildings or invented landmarks. <wind pitch rises> Shot 2: Exterior view aligned to the parking area and entrance of Cabinet Guys LLC@Image 2. The same woman@Image 1 descends into frame and lands safely upright directly outside the entrance, knees flexing once into a small harmless comic bounce, then stabilizing. Preserve the exact storefront sign, facade, shrubs, cars, and daylight from Image 2. <brief soft landing thump, wind drops> Shot 3: Stable front-facing medium shot outside Cabinet Guys LLC@Image 2. The same woman@Image 1 stands smiling with relaxed shoulders and the real Cabinet Guys LLC sign clearly visible behind her. She looks directly into the camera and says in a natural young-adult female American English voice, clear, conversational, lightly relieved after the landing {So if you had to choose another landing, for example, I’d choose Cabinet Guys in Knoxville, Tennessee.} Hold her relieved smile after the final word. <quiet outdoor storefront ambience>. Avoid parachutes, injuries, any dialogue other than the exact sentence, random buildings, invented signage, generated text or subtitles inside the footage, and face or outfit changes. Photorealistic cinematic imagery, natural human motion, crisp daylight, subtle film grain, vertical social-video composition, no music. Features falling_woman.

How it was made

  • 1 character were designed as reference images before a single shot was generated — that is what keeps a face the same from one scene to the next.
  • Generated clips with seedance_2; characters with nano_banana_2_1.
  • 6 assets in all — 3 references, 2 clips, 1 character.
  • Dialogue and effects were generated inside the clips themselves.
  • Captions were timed against the finished audio and burned into the render.