The STORM trend
One photo, one AI image, one AI video, and the whole crowd starts dancing.
Tools in this guide: Topview
Handly Tech · @handlytech The full guide and prompts for my STORM video. One photo of you, one AI image, one AI video.
What you need
- [ ] A clear photo of yourself
- [ ] The original STORM II clip (your reference video)
- [ ] A Topview AI account (GPT Image 2 and Seedance 2.5 are both inside it)
- [ ] The image prompt and video prompt from this page
Step 1: Take a photo of yourself
Step 2: Create the new image with GPT Image 2
1. Open Topview AI and click Image
2. Choose GPT Image 2
3. Upload your photo first (this is Image 1), then the base crowd scene (this is Image 2)
4. Copy the Image prompt below, paste it, and click Generate
💡 The upload order matters. The prompt says Image 1 is you and Image 2 is the scene. If you swap them, the AI gets confused.
🖼️ The new masquerade image (your result from Step 2)
Step 3: Create the video with Seedance 2.5
1. In Topview AI, click Video
2. Choose Seedance 2.5
3. Upload the new image from Step 2 (Image 1), your photo (Image 2), and the original video (muted)
4. Copy the Video prompt below, paste it, and click Generate
💡 Mute the original video before you upload it. The prompt tells Seedance to make a silent video, and you add the music yourself in the next step.
🎬 The original STORM II clip. Open or download it here, then mute it before uploading to Seedance.
Step 4: Add the music and post
1. Put the video in your editor (CapCut, DaVinci Resolve, anything works)
2. Add the STORM song so the dance lines up with the beat
3. Post it and tag @handlytech, I want to see yours 🔥
Image prompt (GPT Image 2)
Use <<<image_2>>> as the strict base image and composition reference.
Use <<<image_1>>> exclusively to preserve the exact identity of the single central man in black.
EDIT ONLY THE SURROUNDING AUDIENCE.
Preserve the following elements from <<<image_2>>> without alteration:
- Strict 4:3 landscape framing
- Single central bald and bearded man
- Central man’s exact face and identity
- His black suit, white shirt, red tie and black gloves
- His position, scale, expression and posture
- His compact black tactical case
- Existing dark architectural background
- Existing cinematic lighting
- Overall crowd size
- Dense staggered formation
- White-and-black color palette
- Camera position and telephoto perspective
Do not regenerate, replace, mask or restyle the central man.
AUDIENCE TRANSFORMATION:
Replace every oversized featureless full-head helmet with an elegant, human-scale luxury masquerade disguise.
The surrounding audience consists of approximately 30–35 distinct adult men and women attending an exclusive secret billionaire masquerade party.
Every guest must have an independent real identity with naturally varied:
- Gender
- Ethnicity
- Age
- Height
- Body build
- Face shape
- Hair color
- Hair length
- Skin tone
The people must clearly look like different real human beings underneath their disguises.
Do not make them identical, cloned, robotic, mannequin-like or perfectly uniform in height.
MASK DESIGN:
Every surrounding guest wears a refined ivory-white sculptural masquerade mask covering approximately the upper half or central portion of the face.
The masks must be proportional to real human faces. They must not enlarge the head.
Allow the following human features to remain naturally visible:
- Hair
- Lower cheeks
- Mouth
- Lips
- Jawline
- Chin
- Some natural skin around the mask edges
Create a sophisticated controlled variety of approximately 10 mask designs:
- Smooth ivory-white Venetian eye masks
- Minimal ivory porcelain domino masks
- Sculpted ivory masks with raised cheek lines
- Thin geometric ivory eye masks
- Refined asymmetrical ivory masks
- Minimal ivory masks with subtle forehead extensions
- Smooth ivory masks with delicate temple wings
- Small ivory cat-eye masks
- Elegant ivory masks with subtle fox contours
- Clean ivory masks with fine architectural ridges
No two adjacent guests may wear the exact same mask design.
All masks remain the same warm ivory-white color with clean porcelain or satin finishes.
Do not add painted patterns, gold decoration, black lace, jewels, glitter, writing or logos.
COUTURE EARS:
Only four guests in the entire crowd may wear masks with subtle integrated couture ears:
- One guest with narrow rabbit ears
- One guest with small cat ears
- One guest with refined fox ears
- One guest with short panther ears
Distribute these four guests across different rows and different sides of the composition.
The ears must be small, narrow, sculptural and elegant.
They must look like expensive avant-garde masquerade headpieces, not animal costumes.
Remove the current oversized rabbit ears.
No additional ears anywhere else.
CLOTHING:
Keep the surrounding audience dressed entirely in coordinated ivory-white formalwear.
Men wear:
- Perfectly tailored ivory-white tuxedo suits
- Black high-neck shirts
- Slim ivory-white ties
- Ivory-white gloves
Women wear:
- Perfectly tailored ivory-white trouser suits
- Structured ivory-white blazers
- Black high-neck shirts
- Slim ivory-white ties or clean collarless styling
- Ivory-white gloves
Keep everyone in suits. Do not introduce dresses, gowns, exposed shoulders or bridal clothing.
Create subtle differences in tailoring appropriate to each person’s body while preserving the appearance of one coordinated group.
POSTURE AND FORMATION:
Preserve the dense staggered formation, but make it naturally human rather than mechanically cloned.
Every person remains:
- Standing completely upright
- Facing directly toward the camera
- Head level
- Shoulders square
- Arms resting naturally downward
- Expression neutral and controlled
- Completely still
Introduce subtle natural height differences between rows and individuals.
Every head should sit between the shoulders of people in the row ahead, following the formation language of the original STORM II scene.
Do not create a perfect military grid.
Do not allow masks, heads or ears to overlap important faces.
Keep the crowd dense around the central man, with only a narrow amount of visual space around his head and shoulders.
VISUAL RESULT:
The audience must feel like wealthy, mysterious and slightly unsettling human guests at a private masked ceremony.
They must not look like robots, aliens, soldiers, mannequins, superheroes or science-fiction clones.
Maintain the powerful white crowd surrounding the single central man in black.
The central man remains the only person with a completely uncovered face, the only black suit and the only red detail.
IMAGE QUALITY:
Preserve the existing high-end cinematic realism and improve the audience with:
- Realistic skin pores
- Natural hair texture
- Distinct human facial anatomy
- Physically realistic porcelain masks
- Accurate mask contact with the skin
- Premium fabric texture
- Sharp individual details
- Controlled cold cinematic lighting
- Natural dimensional shadows
- Subtle atmospheric depth
- Sophisticated psychological-thriller color grading
The final result should resemble a $20-million psychological thriller about a secret masked elite, not a robot army, costume convention or wedding photograph.
ABSOLUTELY NO:
Full-head helmets, featureless oval heads, identical masks, identical people, cloned bodies, robotic audience, mannequin heads, masks larger than human faces, covered mouths, covered chins, covered hair, bald crowd members copied from the central man, oversized rabbit ears, multiple rabbit ears, furry ears, animal heads, gold masks, black lace masks, feathers, horns, antlers, beaks, superhero masks, sci-fi helmets, dresses, wedding styling, additional black suits, additional red elements, changes to the central man, identity drift, distorted anatomy, text, symbols, logos or watermarks.
OUTPUT:
Preserve the original strict 4:3 landscape composition and resolution.
Change only the surrounding audience masks, identities, proportions and natural height variation.
Video prompt (Seedance 2.5)
Use @Video 1, the single uploaded muted video, as the reference for the entrance, camera progression and each dancer’s individual choreography. Use @Image 1 for the scene’s appearance and @Image 2 exclusively for the central man’s identity.
Preserve the reference duration. Output in 4:3 landscape.
ENTRANCE:
Begin with the central man close to the camera, his BACK facing the camera and his face directed toward the crowd. Clearly show his bald head, rear barcode tattoo, small “47” and clean scalp bandage from @Image 2.
He walks AWAY FROM THE CAMERA into the middle of the waiting crowd, following the original performer’s route and entrance timing. He ascends any steps along that route. He never enters from above the crowd or walks down toward the camera.
After reaching his position, he turns to face the camera. Preserve the continuous approach and original camera progression. @Image 1 provides the eventual group appearance, not the opening pose.
STRICT CROWD TIMING: NO DANCING BEFORE 00:15:
From 00:00 until 00:14.999, EVERY surrounding guest remains standing still, facing forward with arms relaxed.
No early head tilts, shoulder rolls, swaying, arm gestures, stepping, bending or anticipatory dance movements. Only subtle breathing and natural blinking are allowed.
This rule applies to every person, including guests at the edges and in the back rows. It overrides any small preparatory movements in the reference.
The first crowd dance movement begins at exactly 00:15.000. Do not start early or delay this onset.
CENTRAL MAN BEFORE 00:15:
After completing his entrance and turning toward the camera, he calmly straightens his red tie once with his right gloved hand.
He then retrieves one compact matte-black flip phone with that same hand, opens it at lower-chest height, briefly checks the screen, closes it and returns it fully to his jacket pocket.
Complete the tie adjustment and phone check BEFORE 00:15. His left hand holds the closed black case beside his leg throughout these actions.
Keep the gestures economical and his expression composed. Do not prolong the entrance or move the dance start to accommodate them.
CENTRAL MAN AFTER 00:15:
As the surrounding crowd starts dancing, he performs one calm case check:
He lifts the compact case to waist height using both hands, supports its base securely with his left hand and forearm, and opens the lid slightly with his right hand. He glances inside, gives one small approving nod, closes and secures the case, then lowers it beside his left leg.
Angle the opening toward him so the contents remain hidden. No visible weapon, light or effect inside.
The phone is already pocketed before this action begins. Never show him handling the phone and opening the case simultaneously.
Perform the case check once, then remain composed while the crowd continues. No dancing, smoking, speaking or exaggerated reactions.
CROWD MOVEMENT AFTER 00:15:
Reproduce the reference dancer by dancer. Match each guest to a distinct performer at the corresponding position in @Video 1.
Preserve each performer’s separate movement path, direction, intensity, pauses and timing offsets. Adjacent people must not automatically share the same pose or movement.
Retain the reference’s overlapping movement: different dancers bending, leaning, turning and moving their heads or shoulders in different directions at the same moment. Preserve those differences wherever they occur in the source.
Do not animate the entire crowd with one shared routine. Do not make whole rows mirror one another. Do not repeat one dance loop across everyone.
Keep the reference’s overall coordination while preserving individual variation. Reproduce moments of unison only where the source actually contains them. Do not invent unrelated random dancing.
APPEARANCE AND CONTINUITY:
Preserve @Image 1’s ivory-white suits, partial ivory masks, distinct guests, dark setting and cinematic lighting.
Preserve the central man’s exact face, beard, bald head, black suit, white shirt, red tie, gloves, barcode and bandage from @Image 2. Apply his identity to him alone.
Keep every guest, mask and prop consistent. The phone and case remain solid objects with realistic hand contact. No duplicated people, changing faces, floating masks, fused bodies, sliding feet or disappearing props.
Follow @Video 1’s camera progression without adding angles, cuts, zooms or transitions.
Generate completely silent video. No music, dialogue, sound effects, captions, logos or watermarks.
Want more AI tricks like this? Follow @handlytech on Instagram, TikTok and YouTube. More free guides: linktr.ee/handlytech
More guides
All guides · All AI tools · Prompts · HandlyTech
