AI Video Generator

Cloud Cabin

Model

Prompt

SCENE CONTEXT

STATIC HANDHELD WIDE SHOT — 14 SECONDS — ONE CONTINUOUS TAKE, NO CUTS.

A small single-storey wooden cabin soars fast through an endless sea of clouds at roughly 120 km/h, like a ship cutting across an ocean, at soft pastel dusk. A young woman leans relaxed against its terrace railing, trailing her hand through thin passing wisps of mist, while far across the cloud sea a massive mythical eagle approaches, lands on the roof above her, and lowers its head down to greet her.

ACTIVE REFERENCES

<<<image_1>>> (face photo) — controls the main character’s face, hair, and body proportions exactly as shown in the uploaded face reference. If the uploaded face photo is female, keep the character as a young woman wearing a high-collared long-sleeve fitted blue dress with its draped gathered skirt, black tights, and brown lace-up boots. If the uploaded face photo is male, keep the character as a young man wearing the masculine version of the same outfit: a high-collared long-sleeve fitted blue garment with the same blue color and fabric character, long sleeves, draped lower construction, dark trousers, and brown lace-up boots. If the uploaded face photo is male, interpret all mentions of “young woman,” “she,” and “her” below as “young man,” “he,” and “him,” while keeping every action and timing exactly the same. 100% matches reference identity.

<<<image_2>>> (creature photo) — controls the creature exactly as previously locked. If the uploaded reference is a bird, it is a photorealistic giant raptor, roughly 4 meters across, cool grey-blue plumage deepening to dark blue-black at the wingtips, pale banded stripe pattern on the upper wing, soft white gradient in the leg/belly down, large piercing yellow eyes, black hooked beak and talons. If the uploaded reference is a dragon, it controls the dragon exactly as shown in the uploaded reference — preserve its anatomy, head shape, wings, texture, color, and overall design exactly, while keeping the same massive scale, calm deliberate approach, landing behavior, lowered-head greeting behavior, and emotional tone in the scene. If the uploaded reference is a dragon, interpret all mentions of “eagle” below as referring to that dragon, while keeping all other behavior the same.

The cabin and terrace are exactly as previously locked: a weathered timber cabin with a mossy roofline, small bells hanging from the eave, dried herbs and small framed charms on the wall, a wooden table with a teapot and book, an old railing overgrown with vines, all suspended above an endless pastel pink-blue cloud sea with a pale, soft-white sun low near the horizon.

IDENTITY LOCK (NON-NEGOTIABLE)

<<<image_1>>>’s face, hair, and body proportions stay exactly as in the reference throughout. She never shows bare arms — the long sleeves of the blue outfit cover her wrists at all times. Her expression stays calm, relaxed, and naturally alive throughout — never frozen or staring — except for the brief startle and the relieved smile described below.

CREATURE SCALE LOCK (NON-NEGOTIABLE)

<<<image_2>>> is genuinely massive — roughly 4 meters across, dramatically larger than the woman and larger than the cabin's roofline is tall. Its wingspan extends well beyond the width of the roof when spread. It is never rendered hawk-sized or falcon-sized; its presence should feel like a small building settling onto the roof.

CAMERA — LOCKED WIDE, OBSERVER ONLY

One fixed wide shot from a single position in the open air off to the side of the house, roughly level with the terrace. The entire cabin, its terrace, and the surrounding cloudscape stay in frame from the first frame to the last. Wide-angle lens, deep focus, everything sharp. The camera travels at exactly the same speed as the house, so the cabin sits locked in place in frame while the clouds rush past it. The camera NEVER moves closer or further away — no zoom, no push-in, no pull-out, no dolly, no crane, no orbit, no pan, no tilt, no rotation, no tracking, no rack focus, no reframing. The distance between camera and house is identical at 0s and 14s.

The ONLY camera movement is HANDHELD SHAKE, and it is noticeable: irregular organic wobble, small sudden bumps and corrections, the operator drifting and pulling the frame back, wind buffeting the camera. Loose, unstable, imperfect, alive — a real human hand fighting the airstream. Never smooth, never stabilized, never gimbal-locked. The shake wobbles around a fixed framing and never turns into an actual move, pan, or change of distance.

HER POSE — RELAXED AND ALIVE

She leans loosely against the railing, forearms and one elbow resting on the worn wooden top rail, her body weight settled and unhurried. One hand is propped gently under her chin, cheek resting lightly against her knuckles, head tilted slightly as she gazes dreamily out at the view — completely relaxed, no tension in her shoulders or posture. Her other arm extends out past the railing, hand open and relaxed, fingers trailing slowly through the air at railing height.
She blinks naturally and unhurriedly throughout the entire shot — soft, ordinary blinks at irregular intervals, never a frozen or staring expression — while the wind occasionally lifts strands of her hair across her face, which she does not fuss over.

COSTUME AND WIND

The blue outfit's long sleeves and gathered, draped lower fabric ripple and swing continuously in the steady headwind, the fabric’s gathered folds catching the wind and shifting with real cloth weight. Her dark bob hair lifts and blows across her face repeatedly.

CLOUD WISPS AT RAILING HEIGHT

The main dense cloud sea sits far below the terrace and does not reach her hand. What she is actually touching are a few thin, barely perceptible wisps and streamers of mist drifting past at exactly railing height — faint, translucent tendrils, almost more air than visible cloud. Because the house is moving forward while these thin wisps hang largely still in the air, they appear to stream backward past her fingers as the cabin passes through them, giving the sensation of trailing her hand through water even though the true cloud floor is far beneath the terrace.

FISH SILHOUETTES IN THE CLOUDS

Far beyond the terrace, faint, indistinct dark silhouettes of fish leap through the surface of the cloud sea exactly as fish would leap from water — arcing up, catching a moment of light along their side, and diving back beneath the cloud surface, small ripples of displaced cloud marking each entry point. They appear only as soft-edged silhouettes, never sharply detailed, in small loose schools of two or three, jumping one after another rather than simultaneously, scattered at varying distances in the background.

BIRD APPROACH — VISIBLE FROM FAR DISTANCE

Early in the shot, the massive eagle is already visible as a distant silhouette far out over the cloud sea, near where the low sun sits, flying on a steady course toward the cabin. It grows steadily larger in frame as it closes the distance, wings beating in slow, powerful strokes, never rushed or erratic — a calm, deliberate approach, not a sudden appearance. It keeps this same altitude as the terrace, tracking straight toward the roof.

ACTION TIMING (within the unchanging frame, calm and unhurried throughout)

0–6s — The cabin flies on through the clouds. She leans relaxed against the railing exactly as described above — chin propped in one hand, the other hand trailing slowly through the thin passing wisps of mist at railing height, blinking naturally, entirely at peace, watching the view and the distant leaping fish silhouettes. Meanwhile, far across the cloud sea near the low sun, the massive eagle appears as a distant silhouette and steadily approaches the cabin, growing larger in frame over these seconds.

6s — The eagle reaches the roof; the entire house suddenly shudders and rocks as it grips the roof ridge with its talons. She flinches — a small, sharp, involuntary startle — and snaps her head upward to look toward the roofline.

6–9s — Its enormous wings beat several times, slow and powerful, to stabilize its balance as it settles its full weight onto the structure. It then bends its long neck downward and back, lowering its head below the eave line in an inverted orientation — the crown of its head toward the ground, its beak and eye angled back up toward her, so it ends up peering at her almost upside-down despite its body remaining perched above the roof. Its wings give a few more stabilizing beats as it holds this lowered-head position. As its head comes level with her, it releases one deep, resonant eagle call — a greeting to a friend, not a threat display. Two or three loose feathers dislodge from the impact of its landing and drift down past the camera in the foreground.

9–11s — Her startled expression dissolves into a warm, relieved smile as she recognizes the huge bird — no more fear, only gladness.
11–14s — She reaches up and gently strokes the lowered head and down the back of its neck with her hand; the eagle leans into her touch, visibly relaxed, its eyes half-closing, nuzzling softly against her palm and fingers, completely at ease with her. The cabin keeps flying, clouds streaming past, and the shot simply holds.

BEYOND THE TERRACE

A dense volumetric cumulus field stretching to the horizon in every direction — no ground, no land, no city, no mountains, no sea, no other structures.

LIGHTING

Soft, diffused dusk light: a pale, near-white low sun near the horizon, its direct touch giving only a soft golden sheen to the nearest cloud tops, while the wider sky and cloud sea read in cooler dusty pink and pale blue-lavender tones. Long, soft, cool-tinted shadows rather than warm amber ones. Gentle, subtle lens flare and haze bloom where the pale sun catches the frame edge, light wrapping softly around her hair and the eagle's feathers as it approaches and descends.

FILM LOOK

Shot on 35mm motion picture film. Visible organic film grain, fine and alive, slightly heavier in shadow. Gentle halation around the brightest edges, soft highlight rolloff, muted film color science with cool-toned shadows, subtle corner vignetting, faint chromatic aberration at the frame edges, slight softness in the extreme corners from vintage glass. Natural 24fps motion blur — authentic filmic texture, not clean digital sharpness.

AUDIO — DIEGETIC ONLY, NO MUSIC

Steady rush of wind at speed, fabric snapping and rippling, creak of old timber, soft rattle of the metal railing and hanging bells, leaves and dried herbs rustling, distant powerful wingbeats growing louder as the eagle approaches, the low deep resonance of its greeting call, the sudden structural creak/thud of the house shuddering under its landing weight, small feathers fluttering past the microphone, faint wind buffeting the camera itself. No music, no score.

STYLE

Photorealistic, cinematic, wide-angle, deep focus, realistic textures of aged wood, cotton/linen dress fabric, skin, hair, and feathers, natural fabric and feather physics, correct real-world scale for both the woman and the massively oversized eagle. Indistinguishable from real footage — documentary footage of an impossible thing, shot on film.

NEGATIVE CONSTRAINTS (known failure modes only)

Smooth stabilized camera, gimbal, steadicam, tripod-locked, motionless frame, no shake.

Any zoom, push-in, pull-out, dolly, crane, orbit, pan, tilt, rotation, tracking, rack focus, or reframing of any kind; subject growing or shrinking in frame from camera movement.

Bare arms, short sleeves, sleeveless, rolled-up sleeves.

Frozen or staring face, no blinking, exaggerated fear/recoiling/stepping back at the bird's arrival — only a brief small startle, then relief and warmth.

Small hawk-sized or falcon-sized bird — the eagle must always read as roughly 4 meters and dramatically larger than the woman and the roofline.

Bird appearing suddenly with no visible approach — it must be seen flying in from the distance before it lands.

Bird landing normally/perched upright facing forward — its head must specifically lower down inverted below the eave to meet her at eye level.

Music, soundtrack, or score of any kind.

Visible ground, land, city, mountains, or sea beneath the clouds; supports or pillars under the house; second storey, upper balcony, or stairs on the cabin.

Clean digital look, oversharpened image, heavy artificial grain overlay, blown-out flares, cartoonish or CGI look, plastic skin, text, watermarks, logos.

12510 / 50000

Reference Images (Optional)

Aspect Ratio

Duration

Resolution

Generate Audio2x creditsBasic+

Required Credits50

Video Preview