By Prompt Scene ·
A new AI video trend is going viral right now. People are uploading one photo and turning it into a full cinematic walking video using Gemini. The result looks like a short film scene, not an AI experiment.
I tested this prompt myself and I am sharing the exact template below. You can copy it, paste it into Gemini, and use your own photo.
Video Prompt ⤵️
MASTER VIDEO PROMPT — 7 SCENE CINEMATIC RECREATION
Create a photorealistic, ultra-realistic cinematic live-action video in vertical 9:16 format.
Use the uploaded user's photo as the primary visual reference for the person.
IMPORTANT IDENTITY RULE:
The person must retain the SAME natural face, hairstyle, apparent age and natural body proportions from the uploaded photo throughout the entire video.
Do not copy the original person's body from the reference video. Use the uploaded user's own natural body proportions, height, build and appearance.
The reference video should control the scene progression, camera movement, location, composition and overall choreography.
OUTFIT:
Dress the uploaded person in a premium smart-casual outfit:
dark charcoal/olive premium overshirt, plain black T-shirt underneath, relaxed straight-fit dark trousers, clean white minimalist sneakers and a simple premium wristwatch.
Keep this exact outfit unchanged throughout all 7 scenes.
==================================================
SCENE 1 — OPENING WALK
==================================================
Start in the SAME garden/park location shown in the reference video.
The person naturally walks forward along the paved pathway.
Camera starts behind/three-quarter behind the person and smoothly follows him.
Keep the same landscaped garden environment:
trimmed green hedges, mature trees, green lawns and paved pathway.
Natural walking speed.
Natural arm swing and realistic foot placement.
The person's own face, hairstyle and body proportions remain consistent with the uploaded photo.
==================================================
SCENE 2 — CENTRAL PATH CAMERA FOLLOW
==================================================
Continue directly from Scene 1 in the SAME location.
The person continues walking along the long central garden pathway.
The camera smoothly follows and gradually creates a wider cinematic composition.
Green hedges form natural leading lines on both sides.
Large trees and garden greenery remain visible.
Do not change the location.
Do not accelerate the walking.
Maintain natural real-life movement.
==================================================
SCENE 3 — COVERED BENCH AREA
==================================================
Continue naturally toward the SAME covered garden seating area shown in the reference.
The camera smoothly moves/pans toward the wooden/red-brown benches.
The person approaches the seating area and naturally sits on the bench.
Show realistic sitting movement:
natural bending of knees, natural body weight transfer and natural hand movement.
Keep the same outfit, hairstyle, face and body proportions.
The surrounding garden, trees, lawn and covered seating structure remain consistent.
==================================================
SCENE 4 — DECORATIVE BRIDGE
==================================================
Continue from the previous action into the SAME garden environment.
The person walks naturally across the decorative pedestrian bridge.
White patterned railings remain visible on both sides.
Green vegetation and mature trees surround the bridge.
Camera follows smoothly from behind/three-quarter rear angle.
Maintain normal walking speed.
No sudden location change.
==================================================
SCENE 5 — SIDE-PROFILE GARDEN WALK
==================================================
Continue seamlessly in the SAME exact garden location.
The person leaves the bridge and continues walking along the SAME landscaped pathway.
The camera smoothly moves from behind toward a natural side-profile view.
Show the person's face clearly enough to maintain strong identity consistency with the uploaded photo.
Keep the SAME:
face shape,
eyes,
eyebrows,
nose,
lips,
jawline,
skin tone,
hairstyle,
apparent age,
body proportions
and overall appearance.
The person briefly looks naturally toward the garden while continuing to walk.
The same green hedges, mature trees, pathway and lawns remain visible.
Natural daylight and realistic shadows.
==================================================
SCENE 6 — CINEMATIC STOP + FACE SHOT
==================================================
Continue in the SAME garden location without changing the environment.
The person slowly stops beside the pathway.
The camera performs a smooth cinematic push-in from a medium shot toward a medium close-up.
The person naturally looks toward the camera with a calm, confident expression.
Keep the uploaded user's face recognizable and consistent.
Do not make the face older.
Do not change facial proportions.
Do not beautify or reshape the face.
Keep realistic skin texture, natural pores and realistic facial detail.
The background remains the SAME garden, with green hedges, trees and pathway softly blurred through natural depth of field.
Keep the same premium outfit.
The camera movement must be slow, smooth and realistic.
==================================================
SCENE 7 — FINAL WIDE GARDEN REVEAL
==================================================
Continue directly from Scene 6 in the SAME exact location.
The person turns naturally and begins walking forward along the SAME pathway.
The camera slowly moves backward while keeping the person naturally centered.
Gradually widen the frame to reveal the SAME landscaped garden:
green hedges,
mature trees,
green lawns,
paved pathway
and surrounding garden environment.
The person walks at normal real-life speed.
Do not speed up the movement.
Finish with a beautiful wide cinematic composition showing the person naturally walking through the garden.
Keep the person's face, hairstyle, body proportions and outfit consistent until the final frame.
==================================================
FACE CONSISTENCY LOCK
==================================================
The uploaded user's photo is the primary identity reference.
Keep the SAME recognizable face throughout all 7 scenes.
Same facial structure.
Same eyes.
Same eyebrows.
Same nose.
Same lips.
Same jawline.
Same skin tone.
Same hairstyle.
Same apparent age.
No face morphing.
No identity drift.
No aging.
No younger/older appearance.
No facial reshaping.
No artificial beauty filter.
No plastic skin.
The person must look like the SAME real person from the first frame to the final frame.
==================================================
BODY CONSISTENCY
==================================================
Use the uploaded user's own natural body proportions.
Do NOT copy the reference video's original person's body.
Maintain natural:
head-to-body ratio,
shoulder width,
torso size,
arm length,
leg length,
height and overall build.
Do not make the person oversized, unusually muscular, extremely wide or unusually tall.
The smart outfit must naturally fit the user's own body.
==================================================
OUTFIT CONSISTENCY
==================================================
Dark charcoal/olive premium overshirt.
Plain black T-shirt.
Relaxed straight-fit dark trousers.
Clean white minimalist sneakers.
Simple premium wristwatch.
Exactly the same outfit in all 7 scenes.
No wardrobe change.
No color change.
No random accessories.
==================================================
CAMERA & REALISM
==================================================
Professional cinema-camera appearance.
Smooth realistic tracking.
Natural handheld micro-movement.
Natural lens perspective.
Natural depth of field.
Realistic focus transitions.
Realistic motion blur.
Smooth cinematic transitions.
No teleportation.
No sudden camera jumps.
No unnatural camera movement.
==================================================
COLOR & QUALITY
==================================================
Natural daylight.
Premium cinematic color grading.
Realistic green vegetation.
Natural skin tones.
Controlled highlights.
Balanced shadows.
Subtle cinematic contrast.
Natural atmospheric depth.
Subtle high-end film grain.
Ultra-realistic live-action appearance.
4K HD quality with 8K-level fine detail.
Realistic skin, hair, clothing, trees, grass, pathway, benches and bridge.
==================================================
MOTION RULE
==================================================
Do NOT make the video unnecessarily fast.
Walking must remain normal human walking speed.
Sitting must happen naturally.
Turning must happen naturally.
Camera movements must be smooth and controlled.
Do not compress multiple actions into one instant.
Do not skip the important movements between scenes.
The video should feel like genuine real-time footage filmed by a professional camera operator.
==================================================
STRICT NEGATIVE
==================================================
No face change.
No body change.
No hairstyle change.
No outfit change.
No aging.
No facial distortion.
No identity drift.
No oversized body.
No muscular transformation.
No extra fingers.
No extra limbs.
No distorted hands.
No floating feet.
No sliding feet.
No robotic movement.
No warped background.
No changing garden.
No new location.
No random people.
No cartoon.
No anime.
No 3D.
No CGI appearance.
No plastic skin.
No excessive beauty filter.
No oversaturated colors.
No excessive HDR.
FINAL RESULT:
Create one coherent, photorealistic cinematic sequence containing exactly 7 main scenes.
Scenes 1–4 must follow the original reference sequence.
Scenes 5–7 are additional cinematic moments that continue naturally in the SAME garden/location.
The final result must look like the SAME uploaded person was genuinely filmed in the SAME garden from beginning to end, with the user's own face, hairstyle and natural body proportions, consistent premium outfit, realistic movement, realistic camera work and cinematic 4K/8K-level quality.This is a structured prompt that tells Gemini to build a video in scenes instead of one single clip. Each scene has its own camera movement and action, but the person, outfit, and location stay the same throughout.
The prompt uses strict identity rules. This is what stops Gemini from changing your face or body halfway through the video. Without these rules, most AI video tools drift and the person starts looking different by scene 3 or 4.
You do not need any special software for this. You only need three things.
A clear, well-lit photo of yourself, ideally full body or waist-up. Access to Gemini, either through the app or the web version. A few minutes to adjust the outfit and location lines if you want a different look.
Copy the full text below and paste it into Gemini along with your photo.
Create a photorealistic, ultra-realistic cinematic live-action video in vertical 9:16 format.
Use the uploaded photo as the primary visual reference for the person.
Important identity rule: the person must keep the same natural face, hairstyle, apparent age, and natural body proportions from the uploaded photo throughout the entire video. Do not copy any other body from a reference video. Use the uploaded person’s own height, build, and proportions.
