AI Video Generator

Hero Object Scene

Model

Prompt

IMAGE 1 = PRIMARY CHARACTER REFERENCE
Use Image 1 as the ONLY authoritative reference for the protagonist.
Preserve the EXACT same person throughout the entire video:
face, facial structure, eyes, skin, hair, hairstyle, body proportions, clothing, shoes, jewelry and accessories.
Do NOT redesign, beautify, age, de-age or reinterpret the person.
IMAGE 2 = LOCATION REFERENCE (OPTIONAL)
Use Image 2 only as inspiration for the environment, architecture, atmosphere, materials and lighting.
Create a NEW cinematic version of the location.
Do NOT copy the exact composition or camera angle.
IMAGE 3 = HERO OBJECT REFERENCE (OPTIONAL)
If provided, Image 3 defines the HERO OBJECT.
Preserve its EXACT:
shape, size, proportions, color, material, texture, construction and design.
If Image 3 is not provided, select ONE recognizable wearable/accessory from Image 1 as the HERO OBJECT.
ABSOLUTE CONTINUITY RULE
IDENTITY AND OBJECT CONSISTENCY ARE NON-NEGOTIABLE.
Every character, object and environment must remain the SAME physical entity throughout the entire video.
The following MUST NEVER change:
* protagonist's face
* hairstyle
* body proportions
* clothing
* HERO OBJECT
* phone
* glasses/accessories
* coffee cup
* coffee lid
* coffee
* location
* architecture
* lighting direction
* time of day
* weather
* materials
* colors
* object dimensions
* object design
A close-up MUST show the SAME physical object from a closer camera position.
Never generate a new version of an object for a macro shot.
Never redesign an object when the camera moves closer.
Never teleport or regenerate objects.
WIDE SHOT → MEDIUM SHOT → MACRO SHOT = THE SAME PHYSICAL OBJECT.
MASTER CONCEPT
Create ONE continuous 12-second, 9:16, ultra-photorealistic LIVE-ACTION cinematic fashion film.
High-budget international fashion campaign aesthetic.
Real human anatomy, practical props, realistic physics, premium cinema optics and high-speed photography.
The central event is an ACCIDENTAL COLLISION followed by BULLET TIME.
The protagonist walks naturally through the environment while:
* RIGHT HAND holds ONE OPEN FLIP PHONE
* LEFT HAND holds ONE TAKEAWAY COFFEE
* HERO OBJECT is naturally worn/carried
Exactly ONE secondary person walks toward the protagonist with exactly ONE Labrador on exactly ONE leash.
No other important characters.
0–3.7s — WALKING REVEAL
Start with a low cinematic close-up of the protagonist's feet.
Camera smoothly cranes upward while tracking backward.
Reveal:
feet → legs → clothing → torso → accessories → face → nearly full body.
The protagonist is naturally talking on the phone.
Maintain exact identity from Image 1.
The environment is cinematic, realistic and spacious enough for the collision.
3.7–4.4s — COLLISION
The secondary person's elbow accidentally clips the protagonist's arm/upper torso.
The collision is small and physically believable.
The protagonist recoils backward.
Three objects physically separate because of the SAME collision:
HERO OBJECT
Receives one short realistic impulse and separates naturally.
iPhone 17 Orange
The SAME phone leaves the RIGHT hand with a short outward/upward impulse.
COFFEE
The SAME cup leaves the LEFT hand with a short outward/downward rotation.
The coffee lid partially opens from the cup's rotation and inertia.
Coffee begins escaping ONLY through the physical opening.
Immediately after release:
RIGHT HAND = EMPTY
LEFT HAND = EMPTY
Hands remain anatomically correct.
4.4–10.4s — BULLET TIME
Time slows to approximately 2–3% of normal speed.
The physical world becomes almost frozen.
OBJECTS STAY IN WORLD SPACE.
CAMERA MOVES THROUGH WORLD SPACE.
The protagonist continues beginning the backward fall.
The HERO OBJECT, coffee cup and phone remain suspended at physically believable positions created by the collision.
The camera first holds a medium-wide view so the viewer understands the entire frozen event.
Then follow this exact camera route:
FULL TABLEAU → GLASSES → COFFEE → iPhone 17 Orange → FULL TABLEAU
HERO OBJECT
Camera physically dollies toward the SAME HERO GLASSES.
Hold a cinematic macro shot.
Perform a subtle 20–30° camera orbit.
Show real materials, texture, edges and manufacturing details.
The object itself remains almost stationary.
COFFEE
Camera physically travels through the frozen scene toward the SAME coffee cup.
Show:
* partially detached lid
* real cup texture
* physical opening
* thin coffee stream
* realistic droplets
* surface tension
* natural reflections
The cup and liquid remain in their existing world-space positions.
PHONE
Camera travels toward the SAME iPhone 17 Orange.
First establish the complete phone.
Then move closer to its existing display.
The phone must remain EXACTLY the same model, color, dimensions, hinge, keypad, screen and open angle.
The screen displays:
BOSS
with a small realistic number underneath.
Do NOT transform or redesign the iphone 17.
RETURN
Camera smoothly dollies backward and reveals the COMPLETE frozen accident again.
All objects remain in their original world-space positions.
10.4–12s — TIME RETURNS
Time suddenly returns to normal speed.
Gravity and momentum resume.
Every object continues from its EXISTING position.
The protagonist completes the backward fall.
The HERO OBJECT falls naturally.
The coffee cup falls and impacts the ground.
The lid reacts naturally and remaining coffee spills through the existing opening.
The SAME phone falls and makes one small realistic bounce.
Objects land at DIFFERENT physically plausible locations.
They do not magically converge.
Camera drops toward ground level and follows the impact.
End with a cinematic low-angle shot showing the fallen protagonist and scattered objects in the SAME environment.
HARD CUT.
PHYSICS & CAMERA RULES
Bullet time means:
FROZEN WORLD + MOVING CAMERA.
Never move objects toward the camera to create close-ups.
Never artificially enlarge objects.
Never teleport objects.
Never regenerate objects.
Every visible change must have a physical cause.
Collision → momentum → object release → bullet-time suspension → gravity resumes → impact.
The entire sequence must feel like ONE continuous practical-effects shot.
CONTINUITY LOCK
The AI MUST maintain the SAME:
PERSON
* face
* eyes
* hair
* body
* clothing
iPhone 17 Orange
* exact model
* exact color
* exact dimensions
* exact hinge
* exact keypad
* exact screen
* exact open angle
HERO OBJECT
* exact design
* exact color
* exact material
* exact dimensions
* exact geometry
GLASSES / ACCESSORIES
* exact same design
* exact same color
* exact same shape
* exact same position until physically released
COFFEE
* same cup
* same lid
* same coffee
* same volume
* same rotation
LOCATION
* same architecture
* same street/environment
* same weather
* same time of day
* same lighting
* same materials
* same colors
* same spatial layout
CRITICAL: NOTHING CHANGES JUST BECAUSE THE CAMERA MOVES.
VISUAL QUALITY
Ultra-photorealistic LIVE ACTION.
Premium fashion campaign.
ARRI Alexa 35 look.
Premium cinema lenses.
Natural skin texture. Realistic hair. Real practical objects.
Realistic reflections. Realistic liquid physics. Real gravity and inertia.
Natural depth of field. Shallow macro depth of field. Subtle 35mm film texture.
Cinematic editorial color grade. 4K Ultra HD.
Maximum facial detail. Maximum material detail. Stable identity. Stable geometry.
Stable environment. Stable object continuity.
NO CGI LOOK.
NO OBJECT DUPLICATION.
NO FACE MORPHING.
NO IDENTITY DRIFT.
NO OBJECT REDESIGN.
NO LOCATION CHANGE.
NO TELEPORTATION.
NO PROP REGENERATION.

7520 / 50000

Reference Images (Optional)

Aspect Ratio

Duration

Resolution

Generate Audio2x creditsBasic+

Required Credits50

Video Preview