Dan Kieft
Dan Kieft
300K subscribers
Subscribe
Join NextGenAI on Skool

How To Start Making AI Videos in 2026 (Beginner Guide)

HERE ARE ALLMY PROMPTS

If you're just getting started, this is the one. Every prompt from the video is here: character sheets, locations, props, and the full Seedance 2.5 prompts. Make your assets, generate your first video. And when you do, show me in the NextGenAI community 😊

Tools I used

01

Which platform should you use?

What to look for

  • A platform that gives you multiple video models. New models come out constantly, the one you use today will get replaced
  • Don't lock yourself into a monthly or yearly plan with one model
  • I use Higgsfield. All the image and video models, plus their own stuff like the API, the MCP and Genjutsu
02

How to make your assets

What are assets?

  • Images, videos or audio you drop in the reference tab so the AI uses them in your video
  • Character sheets, location sheets, objects, even outfits
  • Ideation → generate assets → generate videos → edit. No assets, no consistency. Bad assets = bad videos = wasted credits

The workflow

Ideation, generate asset, generate videos, edit and assemble

Types of references

Character sheet, location sheet, objects, outfit sheet
03

Character sheets

Two ways to do it

  • A character sheet is a blueprint of your character: front, back, close-up
  • Cheapest way: Cinema Studio → Image → AI Cast. One line, 2K, done. Dirt cheap
  • More control: the prompt template + ChatGPT. Describe your character, it writes the prompt
  • Images are cheap, videos are expensive. Generate a lot of images until it's exactly right

AI Cast: blonde woman

Where to find it, and the result

Higgsfield AI Cast in the image tabResult: three-panel sheet of a blonde woman in a dark dress

Cinema Studio → Image → AI Cast. Type what your character looks like, 2K, generate.

A 26 year old american woman with blonde high ponytail, thin build, wearing a simple dark dress and high heels. she wear simple golden accessories

AI Cast: clown

A male clown wearing a colorful vintage costume with a ruffled collar, oversized polka-dot buttons, striped sleeves, baggy trousers, and large red shoes.

His white-painted face features exaggerated black eyebrows, rosy cheeks, a round bright red nose, and a wide painted smile. He has a playful, slightly mischievous expression.

Character sheet prompt template

Paste this into ChatGPT, then tell it who your character is. It fills in [Character] and [Outfit] for you.

A three-panel character reference sheet composed as one horizontal image, divided into three equal vertical panels side by side with thin clean separations. The same person in the same outfit, rendered identically in all three panels. No text, no labels, no numbering anywhere in the image.

The face comes from the attached reference and must match it exactly in every panel: same face shape, bone structure, eyes, brows, nose, lips, skin tone and hair.

CHARACTER: [build and height, body type, hair color, length and styling, makeup, piercings, tattoos, nails, overall vibe]

OUTFIT: [full outfit head to toe: top, bottom, shoes, jewelry, accessories, with colors, fabrics and fit]

LEFT PANEL — full body front view. The figure stands squared to the camera, framed from above the head down to the shoes, arms relaxed at the sides, weight even on both feet. The full outfit is clearly readable.
CENTER PANEL — full body back view. The complete figure from above the head down to the shoes, seen directly from behind, standing straight. Hair, back of the outfit and shoes are clearly readable.

RIGHT PANEL — tight chest-up face close-up. Framed from just above the head down to the collarbones, the face filling the panel, looking straight at camera, neutral relaxed expression, lips closed. Maximum facial detail.

Skin tone is identical in every panel, never darker or tanned on the back view.

The background is one flat 18% neutral gray field, identical in all three panels, with no gradient, no floor, no wall and no vignette. Flat shadowless frontal lighting, even on both sides of the face, no cast shadow under the feet and no shadow on the background. Realistic skin with fine soft pore texture and a matte finish, real hair strands, real fabric texture, never plastic, never CGI.
Even sharpness everywhere, no blur. Photographed on a 50mm prime, soft natural film grain. Photographed not generated.

What I said (Wispr Flow)

Edit my prompt formula in the [Character] and [Outfit] with this context and return to me the full prompt. I want my character to be fashion model. 23-year-old East Asian woman with a short, tousled black bob, wispy curtain bangs, warm fair skin, almond-shaped dark eyes, soft natural makeup, and bold red lipstick. Wearing an oversized charcoal-black tailored blazer with a deep V neckline, statement gold drop earrings. Effortlessly sophisticated editorial aesthetic.

The prompt I got back: fashion model

Result: East Asian fashion model in a charcoal blazer, three panels

Seedream 5.0 Pro, 16:9, 2K. Same prompt in Nano Banana Pro, Nano Banana 2.1 and GPT 2.5 Flare all look a bit different. Pick the one you like, it's 1 to 5 credits a go.

A three-panel character reference sheet composed as one horizontal image, divided into three equal vertical panels side by side with thin clean separations. The same person in the same outfit, rendered identically in all three panels. No text, no labels, no numbering anywhere in the image.
The face comes from the attached reference and must match it exactly in every panel: same face shape, bone structure, eyes, brows, nose, lips, skin tone and hair.
CHARACTER: A 23-year-old East Asian female fashion model, approximately 175 cm tall, with a slender, statuesque runway-model physique, elegant proportions, long legs, narrow waist and relaxed, poised posture. Warm fair skin, almond-shaped dark brown eyes, delicate facial features and naturally defined cheekbones. Short, tousled jet-black bob haircut with soft, wispy curtain bangs framing her face, slightly undone with natural volume and individual loose strands. Soft natural makeup with subtly defined eyes, lightly sculpted cheeks and bold matte red lipstick. Small pierced ears adorned with statement gold drop earrings. Clean, well-groomed natural nails, no visible tattoos. Effortlessly sophisticated, confident high-fashion editorial aesthetic with understated elegance.
OUTFIT: An oversized charcoal-black tailored blazer with structured shoulders, wide notched lapels, a plunging deep V neckline and a relaxed, slightly boxy silhouette. The blazer is worn as a statement piece with no visible inner top, featuring luxurious matte wool-blend suiting fabric, subtle texture and precise tailoring. Paired with matching charcoal-black high-waisted tailored trousers, featuring a relaxed straight-leg cut, soft front pleats and a full-length drape. Black pointed-toe leather ankle boots with slim heels, partially visible beneath the trousers. Statement sculptural gold drop earrings with a polished metallic finish, providing a striking contrast against the dark tailoring and black hair. No necklace, no handbag, no additional accessories. A refined, minimalist luxury fashion look with strong editorial styling and a monochromatic silhouette.
LEFT PANEL — full body front view. The figure stands squared to the camera, framed from above the head down to the shoes, arms relaxed at the sides, weight even on both feet. The full outfit is clearly readable.
CENTER PANEL — full body back view. The complete figure from above the head down to the shoes, seen directly from behind, standing straight. Hair, back of the outfit and shoes are clearly readable.
RIGHT PANEL — tight chest-up face close-up. Framed from just above the head down to the collarbones, the face filling the panel, looking straight at camera, neutral relaxed expression, lips closed. Maximum facial detail.
Skin tone is identical in every panel, never darker or tanned on the back view.
The background is one flat 18% neutral gray field, identical in all three panels, with no gradient, no floor, no wall and no vignette. Flat shadowless frontal lighting, even on both sides of the face, no cast shadow under the feet and no shadow on the background. Realistic skin with fine soft pore texture and a matte finish, real hair strands, real fabric texture, never plastic, never CGI.
Even sharpness everywhere, no blur. Photographed on a 50mm prime, soft natural film grain. Photographed not generated.

Character sheet of yourself: what to add

Dan's selfie reference

Already have a character, or want yourself in the video? Attach the image, paste the same template, put this one line in front of it.

Generate the final prompt based on my reference

The prompt I got back: me

Result: Dan in a black sweatshirt, three panels

When you generate, add the selfie as @image1 so it actually uses it. This is the sheet I use whenever I'm in a video.

A three-panel character reference sheet composed as one horizontal 16:9 image, divided into three equal vertical panels side by side with thin clean separations. The same person in the exact same outfit, rendered identically in all three panels. No text, no labels, no numbering anywhere in the image.

The face comes from @image1 and must match the reference exactly in every panel: identical face shape, bone structure, forehead, hairline, eyes, eyebrows, nose, lips, ears, jawline, skin tone, mustache, facial hair and hairstyle. Preserve the person's unique facial identity and natural asymmetries without beautification or alteration. Use the reference exclusively for facial identity, hairstyle, facial hair and visible upper-body appearance. Do not reproduce the original background.

CHARACTER: A young adult man of average height with a naturally slim-to-average build, lean proportions and relaxed posture. Fair, light skin with subtle warm peach undertones and natural pinkness around the cheeks and nose. An elongated oval face with a gently defined jawline, subtle cheekbones, a medium straight nose, naturally pink lips, hazel-brown eyes and moderately thick brown eyebrows. Short dark-blond to light golden-brown hair, with slightly darker roots and sides, a naturally voluminous textured top, casually swept to one side with a few loose strands and slightly tousled styling. A distinct natural light-to-medium brown mustache across the upper lip, with fine, sparse stubble along the chin and jawline. Clean, natural skin with visible subtle pores and minor skin imperfections. No makeup, no visible piercings, no visible tattoos, short natural fingernails. An approachable, understated, casual appearance. Maintain the exact facial proportions, hair volume, facial hair density and natural skin characteristics from @image1.

OUTFIT: A plain dark charcoal-black crewneck sweatshirt matching the reference image, made from soft matte cotton fleece with a naturally relaxed fit, a round ribbed neckline, full-length sleeves and ribbed cuffs. No graphics, logos or visible branding. Paired with dark charcoal straight-leg trousers with a comfortable, slightly relaxed fit, subtle natural fabric folds and a clean waistband. Minimal white leather low-top sneakers with white laces, simple white soles and no visible branding. No jewelry, watch, glasses, hat or additional accessories. Clothing must remain identical in color, material, fit and construction across every panel.

LEFT PANEL — full body front view. The figure stands squared to the camera, framed from above the head down to the shoes, arms relaxed at the sides, weight even on both feet. The full outfit is clearly readable. Natural body proportions, realistic posture, shoulders relaxed, hands visible and fingers naturally separated. Face looking directly toward camera with a neutral relaxed expression.

CENTER PANEL — full body back view. The complete figure from above the head down to the shoes, seen directly from behind, standing straight. The back of the hairstyle, natural hair growth pattern, sweatshirt seams, fabric drape, trousers and sneakers are clearly readable.
Preserve the same physique, shoulder width, height, skin tone, clothing proportions and hair color as the front view. No head rotation or visible facial features.

RIGHT PANEL — tight chest-up face close-up. Framed from just above the head down to the collarbones, the face filling the panel, looking straight at camera, neutral relaxed expression, lips closed. Maximum facial detail. Match @image1 with particular attention to the precise eye shape and spacing, iris color, natural eyebrow thickness, nose bridge and tip, lip shape, mustache pattern, delicate facial stubble, cheek contours, hairline and individual textured strands of golden-brown hair. Preserve subtle facial asymmetry and natural skin texture. No retouching, no smoothing, no exaggerated features.

Skin tone is identical in every panel, never darker or tanned on the back view.

The background is one flat 18% neutral gray field, identical in all three panels, with no gradient, no floor, no wall and no vignette. Flat shadowless frontal lighting, even on both sides of the face, no cast shadow under the feet and no shadow on the background. Realistic skin with fine soft pore texture and a matte finish, real hair strands, real fabric texture, never plastic, never CGI.
Even sharpness everywhere, no blur. Photographed on a 50mm prime, soft natural film grain. Photographed, not generated.
04

Location sheets

Keep your location the same from shot to shot

  • Cinema Studio → Image → Cinematic Locations. Free, and the lighting looks real
  • Say "3/4 wide-angle view". Without it you get a completely different shot, the AI doesn't know what the camera is doing
  • Say "no people", it's a location shot
  • Cinematic Locations isn't always accurate. Try the other models too and pick your favourite

Back alley with a yellow Porsche

Result: graffiti back alley, yellow Porsche
3/4 wide-angle view of an empty back alley in the morning with numerous colorful graffiti and a single bright yellow Porsche in the middle of the alleyway. no people. only a single car
05

Prop sheets

For stuff that has to look the same every time

  • Products, or anything that needs to look a certain way
  • Needs to be accurate? GPT 2.5 Sunburst, GPT 2.5 Flare or Nano Banana 2.1. They understand context
  • Two views, front and top-down
  • Grey background, not white. White overexposes when you use it as a reference

300K subs cake

Result: pink cake with 300K SUBS candles, front and top-down
A celebratory multi layered sponge cake with “300K SUBS” displayed prominently on top using decorative letter-shaped candles. 
Two views of the same object: FRONT view on the left of the frame, and topdown view on the right. 
Both centered, same scale. Flat seamless mid-grey background
06

How to prompt Seedance 2.5

Same structure for text-to-video and reference-to-video

  • Goal → character references → scene → stages (what happens in each shot) → visual style → audio → maintain consistency
  • Sounds like a lot. You don't write it yourself, the skill file does it
  • ChatGPT → Customize → Skills → Add → upload from your computer. Download the skill here. Then just describe your idea

The prompt structure (example: Dan shuts down the AI)

This is what a finished Seedance 2.5 prompt looks like. Every prompt below follows the same shape.

[GOAL]
Create a 15-second cinematic scene where Dan walks through a futuristic AI control room, takes a remote from his pocket, and shuts down the AI system with one click. Build tension during his approach, then finish with a quiet, confident moment.

[CHARACTER AND REFERENCES]

Dan - Dan is a handsome 26-year-old man. Match his face and hair to the attached reference.
He wears a black jacket, plain white T-shirt, dark trousers, and black boots. He moves calmly and confidently.

Location - Use this image for the control room’s architecture, server racks, and central AI machine. Preserve its layout across all shots.

Remote - A small matte-black remote with one circular red button. It starts inside Dan’s right jacket pocket.

[SCENE]

Night inside a large industrial AI control room. A central aisle leads toward the AI machine, with tall server racks on both sides.

Blue screens illuminate the room and reflect across the dark floor. Thin haze catches the light. A warm emergency light above the entrance outlines Dan’s shoulders and remains on after the machine shuts down.

Dan is the only person present.

[STAGE 1 | 0–4 SECONDS | THE APPROACH]

Camera: Medium side-angle tracking shot at waist height. Move smoothly alongside Dan, keeping his upper body and walking direction clearly visible.

Action: Dan walks down the central aisle toward the AI machine. His shoulders are relaxed, his eyes stay on the machine, and his expression is serious. Server racks pass through the foreground.

End transition: Dan approaches his stopping point. Cut while he is still walking to a close-up of his feet, matching the movement.

[STAGE 2 | 4–6 SECONDS | THE FINAL STEPS]

Camera: Low-angle close-up of Dan’s boots. Track beside them, slowing with him.

Action: He takes his final two steps. Blue reflections slide across the leather as each boot meets the floor.
End transition: He stops with both feet firmly planted, facing the machine. Cut to his upper body.

[STAGE 3 | 6–8 SECONDS | THE REACH]

Camera: Medium close-up from Dan’s front-right side, framing his face and right jacket pocket. Slowly push closer.

Action: Keeping his gaze on the machine, Dan lowers his right hand and slides it into his right jacket pocket.

End transition: His hand begins to withdraw. Cut on that movement to a close-up, continuing the same action without repeating it.

[STAGE 4 | 8–15 SECONDS | THE SHUTDOWN]

Camera: Close-up of Dan’s right hand and pocket. Tilt slightly upward as he lifts the remote, then holds still. The AI machine remains visible behind his hand, initially out of focus.

Action: Dan pulls the remote completely out of his pocket and raises it to waist height. His thumb rests on the red button. He says, “Let’s give AI a break,” then presses the button once.

Shift focus from the remote to the machine as its blue screens go dark and its central light fades out. The warm emergency light stays on.

End: Hold the final composition for one second, with the remote in the foreground and the powered-down machine behind it. No internal cuts.

[VISUAL STYLE]

Photorealistic science-fiction thriller. Realistic skin, fabric, metal, and reflections. Cool blue machine light contrasts with the warm entrance light.

Smooth, controlled camera movements, shallow depth of field in close-ups, subtle film grain, and natural motion blur. Keep the atmosphere tense and cinematic.

[AUDIO]

Music:
(Low electronic music builds during Dan’s approach, then stops when he presses the button.)

Sound effects:
<Steady server hum, heavy footsteps, jacket fabric moving, and one crisp button click. After the click, the machinery winds down into quiet room ambience.>

Dialogue: During Stage 4, before pressing the button, Dan speaks in natural English, calmly and with dry humor:
{Let’s give AI a break.}

[MAINTAIN CONSISTENCY]

Keep Dan’s face, outfit, screen direction, room layout, and remote identical across every shot.
Match his body position and hand movement across cuts. The remote comes from his right jacket pocket and stays in his right hand. The machine shuts down only after he presses the button.
07

Text-to-video

Just a prompt, no references

  • Paste it in, pick the duration, generate
  • Don't use the first prompt it gives you. Read it, change what you don't like, go again
  • Generate in Seedance 2.5 draft first. If you like it, upscale later. First tries are rarely the best

Example 1: what I said (Wispr Flow)

I want to make a video that is about 12 seconds long, where we have a boyfriend that is surprising his girlfriend by coming home early from vacation. He walks up to the door and rings the bell. The girlfriend opens it, but she's cheating on him with another man. What happens then is she's like, "Oh, the pizza guy's here," but then she opens the door and sees her boyfriend, and they're both shocked. That's kind of the scene that I want to create.

Example 1: the prompt I got back

[GOAL]
A cinematic, realistic relationship drama: a boyfriend returns early from abroad to surprise his girlfriend with flowers, only to discover her with another man. Three shots, with the preparation, doorbell and doorway reveal captured in one continuous middle shot.

[CONTINUITY AND DOORWAY BLOCKING]
Boyfriend: late twenties, short dark hair, charcoal wool overcoat over travel clothes, visibly wet from the rain. His bouquet is initially protected beneath his coat. His house key stays in his pocket because he chooses to ring the bell.
Girlfriend: late twenties, loose hair, casual thin-strap tank top and lounge trousers.
Other man: late twenties, casual T-shirt and trousers.

Location: a London terraced-house doorstep at night, steady rain, wet brickwork, reflective pavement and warm amber light inside. The front door opens inward. The doorbell is mounted on the exterior wall beside it.

The boyfriend stands outside, one step back from the threshold, facing the door.
The girlfriend personally opens the door from inside. She stands directly at the threshold, within arm’s reach of the handle, and remains there throughout the reveal.
The other man stands immediately beside and slightly behind her, close enough to keep his arm around her waist. Both are right at the entrance, not farther down the hallway.
The door moves only through the girlfriend’s physical action.

[STAGE 1 | 0–2 seconds | Arriving]
Camera: medium side-angle tracking shot, moving parallel to the boyfriend as he runs left to right toward the doorstep.
Action: he hurries through the rain, one forearm holding his coat closed over the hidden bouquet. He reaches the door and stops, breathless but smiling with anticipation.
Transition: hard cut.

[STAGE 2 | 2–9 seconds | Preparation, doorbell and reveal — continuous shot]
Camera: begin in a medium three-quarter side view outside the house, showing the boyfriend’s face, upper body and the exterior doorbell beside the closed door. As he rings it, smoothly arc behind his shoulder, settling into an over-the-shoulder view of the entrance. Frame the doorway closely enough to clearly show the girlfriend opening it and standing at the threshold. The camera stays outside. No internal cuts.

2–5 seconds:
The boyfriend parts his coat and draws out the dry bouquet with his left hand. With his right hand, he quickly smooths his wet hair, straightens his collar and squares his shoulders. He smiles expectantly, then visibly presses the exterior doorbell with his right index finger.

5–6 seconds:
He raises the flowers slightly and waits. The door is still closed. The girlfriend is immediately behind it, close to the interior handle. Her voice comes from just behind the door, muffled. She says casually in British English: {It must be the pizza.}

6–9 seconds:
The handle visibly turns as the girlfriend operates it from inside. She pulls the door inward with her right hand and takes a small step back to clear its swing. As the gap widens, reveal her directly behind the moving door, her right hand visibly gripping the interior handle. She finishes opening it and stays at the threshold, still holding the handle.

She is mid-laugh, her head turned slightly toward the other man immediately beside and behind her. His arm rests around her waist. They are together at the front entrance, filling the doorway.

She turns toward the boyfriend standing directly in front of her. Her smile vanishes; her lips part and her shoulders stiffen. The other man follows her gaze and stops laughing. Her hand freezes on the door handle. Hold briefly on the two of them at the threshold, with the boyfriend’s shoulder and raised bouquet in the foreground.

Transition: hard reverse cut to the boyfriend.

[STAGE 3 | 9–12 seconds | Frozen outside]
Camera: chest-up close-up from just inside the doorway, looking out at the boyfriend. The bouquet is visible at the bottom of frame. Hold steady for the full three seconds.
Action: the faint remnants of his hopeful smile linger, then fade as he understands what he is seeing. His eyes widen slightly, his breath catches and his lips part without words. The bouquet slowly sinks a few centimetres as his arm slackens. Rain runs through his hair and down his face. He stays rooted to the doorstep, staring at his girlfriend, unable to decide what to do. End on his stunned stillness.

[VISUAL STYLE]
Grounded cinematic realism, subtle film grain and restrained, natural performances. Cool rainy exterior contrasted with warm amber interior light. Preserve readable pauses and a sustained final reaction.

[AUDIO]
<Continuous rainfall, hurried wet footsteps, coat and bouquet-paper rustle.>
<The outside button triggers a doorbell chime inside.>
Dialogue language: British English. Girlfriend, immediately behind the closed door: {It must be the pizza.}
<Handle turning, latch releasing, door opening under her hand. Brief laughter becomes clearer as the door opens, then stops at recognition. Rain continues through the final speechless pause.>

[EXCLUSIONS]
The door must not open by itself. The girlfriend must visibly operate and hold the door handle.
Do not reveal the couple standing in the middle or at the far end of the hallway.

[MAINTAIN CONSISTENCY]
Exactly three adults. Preserve their identities, clothing, bouquet ownership and doorway positions. The girlfriend opens the door and stays at the threshold, with the other man immediately beside and behind her. The boyfriend remains outside. Stage 2 is one uninterrupted shot.

Example 2: what I said

The ChatGPT message, dictated

Yes it's messy, that's Wispr Flow. It gave me a shot list first, then I left notes on everything I didn't like, over and over, until I had the final prompt.

use @Seedance 2.5 Prompter to create the prompt

so I want to make a hangover like scene where a guy is a very ordinary guy looks like a nerd he starts off by waking up and then He does not realize that he's very extremely groggy like hungover. He does not realize the state of the apartment is everywhere is messy. There's a there's a Chicken there's a broken aquarium with like crabs on the floor and inside the corner There's a horse. I believe he just reveals a horse right at the background, but he doesn't realize he's a he just wants to go to a bathroom So he he just walks so he cuts a body mounted cat bounty body mounted snorey cam of him just like grow grilly Finding his glasses he puts on it like slowly walks towards a bathroom Then only when he goes to the bathroom he As he's watching space he hears a sound and he saw he sees a huge Alligator is buffed up which really wakes him about for a shot the door

give me the shot list camer angle + movement first for me to ocnfirm 15 sec scene

Example 2: the prompt I got back

[GOAL]
A photorealistic, 25-second morning-after comedy scene in nine connected shots. An extremely groggy, ordinary nerdy man wakes in a wrecked apartment, puts on damaged glasses and shuffles into the bathroom. A sound makes him discover an enormous alligator staring directly at him from a water-filled bathtub. He initially dismisses it and resumes washing before the realization finally hits.

[CONTINUITY]
The same skinny adult man throughout: messy brown hair, clean-shaven face, rumpled gray T-shirt, plaid boxer shorts and bare feet.

His black rectangular glasses initially lie on a low table beside the sofa, surrounded by empty beer cans and bottles. His RIGHT lens is intact. His LEFT lens is shattered, with spiderweb cracks and a small missing section; the remaining glass stays inside the frame. Preserve this exact damage throughout. He wears the glasses crookedly without noticing.

The sofa connects to a short, clear route toward the bathroom. Inside, the sink is beside the doorway; the bathtub is farther to the side. The bathroom door opens inward and remains open until his escape.
Show connected physical actions across every cut. His early movements are slow, heavy and uncoordinated.

[SHOT 1 | 0–3 SECONDS | OVERHEAD WAKE-UP]
Camera: Directly overhead close-up of his face, slowly zooming out to include his upper body.

He lies motionless on the sofa. His eyelids gradually open, close again, then barely reopen. He exhales heavily and slowly extends one arm toward the side table, keeping his head on the cushion.

Transition: Cut to his reaching hand.

[SHOT 2 | 3–5 SECONDS | GLASSES AMONG EMPTY BEERS]
Camera: Static tabletop close-up.

His damaged glasses lie among several empty beer bottles and crushed beer cans, surrounded by bottle caps and dried drink rings. Clearly show the intact right lens and shattered left lens.

His fingertips search blindly between the bottles and cans. He nudges an empty can, making it scrape across the table, then finds the glasses. He loosely grasps a temple arm and lifts them.

Transition: Match cut as the glasses rise toward his face.

[SHOT 3 | 5–8 SECONDS | PUTTING THEM ON]
Camera: Return to the overhead angle, framing his upper body and the sofa edge.

Continue the glasses movement. Still reclining, he slowly slides them onto his face, leaving them crooked. He pauses with his eyes half closed, then laboriously rolls his shoulders forward, sits up and lowers his feet to the floor.

Transition: Cut as his feet settle.

[SHOT 4 | 8–10 SECONDS | APARTMENT REVEAL]
Camera: Low wide angle across the living room, gently pulling back.

Already seated with his feet on the floor, he leans forward, pushes himself upright and begins his first shuffling step.

The widening composition reveals scattered clothes, overturned furniture, a chicken wandering beside the sofa, a broken aquarium with crabs crawling through shallow puddles, and a full-sized horse standing quietly in the rear corner. Keep his walking route clear. He notices none of the chaos.
Transition: Match cut on his first step.

[SHOT 5 | 10–13 SECONDS | BODY-MOUNTED SNORRICAM]
Camera: Body-mounted Snorricam.
Movement: Keep the camera fixed relative to his torso while he moves.
Speed: Match his body motion exactly.
Framing: Keep him close, centered and facing the camera as the background moves around him. His face and upper chest fill the frame, with a constant camera-to-body distance. The rig remains invisible and his hands remain outside the frame.

Continue his shuffling steps toward the bathroom. His face stays anchored in the composition while the apartment sways and bobs behind him. His head droops, his mouth hangs slightly open and his eyelids sag behind the damaged glasses, which are already on.

He reaches the bathroom door. His shoulder shifts forward and the handle clicks offscreen.

Transition: Cut on the handle click.

[SHOT 6 | 13–17 SECONDS | TAP, SOUND, TURN]
Camera: Medium side angle inside the bathroom, showing the doorway and sink. Keep the bathtub outside the frame.

He opens the door, steps inside and moves directly to the nearby sink. He turns on the tap and lowers his hands toward the running water.

Only after the tap starts, a low reptilian rumble and a soft bathwater slosh come from offscreen.
The sound makes him pause. He slowly turns his head toward the bathtub, still looking exhausted rather than frightened.

Transition: Cut along his eyeline.

[SHOT 7 | 17–19 SECONDS | ALLIGATOR STARING DIRECTLY AT HIM]
Camera: His exact point of view from beside the sink, looking slightly down into the bathtub. A subtle forward drift.

An enormous alligator rests in a bathtub full of water. Its heavy body is submerged, with its armored back, eyes and broad snout above the surface.

Its head is turned specifically toward the man’s position at the sink. Its snout points toward him, and its open eyes hold a direct, steady stare into the POV camera, establishing unmistakable eye contact with him. Hold this stare throughout the shot.

Its mouth stays closed. Its body remains almost motionless, with only gentle ripples across the bathwater. No thrashing or active splashing.
Transition: Cut back to the man at the sink.

[SHOT 8 | 19–22 SECONDS | DELAYED REALIZATION]
Camera: Return to the same bathroom side angle in a medium shot.

He slowly turns away from the alligator and back toward the sink, apparently dismissing the impossible sight. He resumes collecting water in his cupped hands and begins lifting it toward his face.

Then his hands stop halfway. His eyes widen behind the damaged glasses. Water drains between his fingers. His breathing catches as he finally processes what he saw. He snaps his gaze back toward the bathtub and begins recoiling toward the doorway.

Transition: Cut during his recoil.

[SHOT 9 | 22–25 SECONDS | PANICKED ESCAPE]
Camera: Static wide shot from the hallway, looking through the open bathroom doorway. Show his route from the sink to the exit.

Continue the same recoil without resetting his position. He screams, scrambles out through the doorway, turns and pulls the bathroom door shut with a violent slam.

End on him pressed against the closed door, breathing hard, eyes wide and completely awake.
His crooked, damaged glasses remain on his face.

[VISUAL STYLE AND PERFORMANCE]
Natural morning light, muted colors, dusty sunlight, realistic skin and believable apartment clutter. Present the animals with complete visual seriousness.

The opening feels slow and physically heavy. His expression stays sleepy through his first look at the alligator. The panic arrives only after he turns away and attempts to resume washing.

[AUDIO]
<Heavy sleepy breathing, cushion creaks, fingertips brushing the tabletop, an empty beer can scraping, faint chicken clucks, dragging bare footsteps, a bathroom handle click and running tap water. A low reptilian rumble and gentle bathwater slosh explicitly trigger his turn. The tap continues through his delayed realization, followed by a sharp human scream, frantic footsteps, a hard door slam and breathless panting.>

[MAINTAIN CONSISTENCY]
Preserve his identity, clothing and damaged glasses: his right lens intact, his left lens shattered. Maintain the apartment layout and match movement across cuts. The alligator rests in the water and looks directly at him during its reveal. Preserve the action order: tap on, sound, turn, eye contact with the alligator, return to washing, realization, escape.
08

Reference-to-video

Most control, this is what I use most

  • Drop all six assets into ChatGPT and describe what each one is
  • In Higgsfield, make sure every reference is connected to the right @image number, and the duration matches your prompt
  • 1080p is the only one worth it if you're actually using the video. Draft if you're not sure
  • Voices sound off? Record your own lines, turn them into black-screen MP4s (not MP3s) and reference them as @Video 1 and @Video 2

The six references

Dan character sheetBruiser character sheetSkinny gangster character sheetBig gangster character sheetBack alley location sheetKinder Surprise prop sheet

What I said

here are all my references @image1 is Dan, 2 is bruiser his friend who always argues with him but follow him anyway, 3 is the skinny gangster, and 4 is the big scary gangster, 5 is the location sheet at a back alley with a yellow car, and 6 is the prop sheet which is the kinder surprise use Seedance 2.5 Prompter to make the prompt. 30 sec scene. it starts with Dan and bruiser carrying a medium sized sack over his shoulder, both entering the alley, Dan feeling normal but bruiser super nervous and question Dan 'are you sure this is the right place?'
Dan doesnt answer and they turn a corner, revealing the gansters waiting for them at the far end of the alleyway, sitting on the car hood and leaning against the yellow car. then cut to a close up, intimating shot maybe a low angle of 2 gangster. cut back to bruiser and dan walking to meet them, bruiser gets more freaked out and panics a bit 'i told you this is a bad idea!"
Dan just tells bruiser to relax and that it's a simple deal.
Dan and bruiser come to a stop at the car, and Dan nods sharply at them once 'you got the good stuff?' the skinny gangster shares a look with the big gangster, which the big gangster looks at them intimidatingly 'you got the money?'
Dan looks at bruiser and bruiser thinks about it for a moment before saying 'whatever!' and drops the sack on to the floor, and then a close up shot of the reveal of $100 bills inside. the gangsters smile when look at money and the skinny gangster pop open the trunk and nods at Dan and bruiser to walk over.
Dan starts walking to the trunk, and bruiser panics a little before following Dan from behind more and more nervous. A low to high angle from the car trunk as Dan head and then bruiser appears into frame looking topdown onto the hidden package.
Dan says smugly, glancing at bruiser, 'i told you this is worth it.
Bruiser is very impressed and stunned "Arent this banned since the 70s? reveal to the trunk stashed with kinder surprises as Dan says 'yes, yes they are'

The prompt I got back

Good, but that's not how I sound, and the gangster sounds like Vin Diesel.

[GOAL]
A 30-second cinematic crime-comedy scene. Dan and his nervous robot friend Bruiser enter a back alley for a suspicious cash deal with two intimidating gangsters. Play the encounter completely straight, building tension until the final reveal: the yellow car’s trunk is packed with Kinder Surprise eggs.

[REFERENCES]
<<<image_1>>> — Dan. Preserve his exact face, tousled light-brown hair, thin mustache, black sweatshirt, black trousers and white sneakers. All panels depict one man.

<<<image_2>>> — Bruiser. Preserve his short, stocky robot body, weathered white-gray metal, rust-orange armor, “07” chest marking, antennas and blue LED eyes. Keep his approximately one-meter height relative to the adults. The panels show different views and expressions of the same robot.

<<<image_3>>> — Skinny gangster. Preserve his face, scrawny build, black beanie, oversized gray hoodie, worn graphic T-shirt, baggy jeans, jewelry and sneakers. All panels depict one man.
<<<image_4>>> — Big gangster. Preserve his face, bald head, facial tattoos, muscular build, black bomber jacket, dark clothing, chain and sneakers. All panels depict one man.

<<<image_5>>> — Location and car. Match the gritty graffiti-covered alley, industrial walls, scattered debris, daylight and yellow sports car. Use this reference for the environment and vehicle only; the two gangsters are defined by <<<image_3>>> and <<<image_4>>>.

<<<image_6>>> — Kinder Surprise packaging. Match the egg shape, crinkled white-and-red foil and colorful Kinder Surprise branding. The two views show the same product design, repeated across many normal-sized eggs in the trunk.

Do not reproduce reference-sheet panels, studio backgrounds or measurement graphics.

[CONTINUITY]
Exactly four characters: Dan, Bruiser, skinny gangster and big gangster.

Bruiser alone carries one medium-sized, loosely cinched canvas sack over his shoulder. Dan walks beside him empty-handed. The sack contains bundles of US $100 bills and remains on the ground beside the car after Bruiser drops it.

The skinny gangster initially sits on the yellow car’s front hood. The big gangster leans against the adjacent front fender. The trunk is at the rear of that same stationary car.

Keep the Kinder Surprise eggs completely hidden from the audience until the final shot. When the trunk opens, frame its exterior and the characters without exposing the contents.

[STAGE 1 | 0–5 seconds | Entering and rounding the corner]
Camera: Medium-wide three-quarter frontal tracking shot, retreating with Dan and Bruiser. As they turn the corner, arc continuously beside them into an over-the-shoulder wide view down the alley. No internal cut.

Initial state: Dan walks casually, shoulders loose. Bruiser keeps close beside him, gripping the sack over his shoulder, blue LED eyes flicking toward the surrounding walls.

Action sequence: Bruiser looks up at Dan.
Dialogue language: English. Bruiser asks anxiously: {Are you sure this is the right place?}
Dan gives no answer and keeps walking. They round the corner, revealing the two gangsters waiting at the far end of the alley beside the yellow car: skinny gangster sitting on the hood, big gangster leaning against the fender.

End state: Dan and Bruiser occupy the foreground; the gangsters and yellow car are clearly visible in the distance.
Transition: Hard cut.

[STAGE 2 | 5–6 seconds | Intimidating gangster insert]
Camera: Tight low-angle two-shot from near hood height, looking upward at both gangsters. Static camera.

Action sequence: The skinny gangster stares toward the approaching pair from the hood. Beside him, the big gangster lowers his chin and fixes them with an unblinking glare. Neither speaks or smiles.
Transition: Hard cut.

[STAGE 3 | 6–10 seconds | Nervous approach]
Camera: Medium frontal tracking two-shot, moving backward as Dan and Bruiser approach the car. Frame low enough to keep the short robot fully readable beside Dan.

Action sequence: Bruiser hunches around the sack, glances toward the gangsters, then urgently looks up at Dan.
Bruiser says in English: {I told you this is a bad idea!}
Dan keeps his eyes ahead and answers with effortless calm.
Dan says in English: {Relax. It’s a simple deal.}

End state: They reach the gangsters and slow to a stop.
Transition: Hard cut.

[STAGE 4 | 10–12 seconds | Dan opens the deal]
Camera: Medium close-up of Dan over the big gangster’s shoulder. Static, eye-level framing.

Action sequence: Dan stops beside the front of the car, meets the gangsters’ eyes and gives one sharp upward nod.
Dan asks in English: {You got the good stuff?}

End state: Dan holds his confident expression.
Transition: Reverse-angle cut.

[STAGE 5 | 12–14 seconds | The money question]
Camera: Tight two-shot of the gangsters from Dan’s position, slightly low angle. Static camera.

Action sequence: The skinny gangster turns his eyes toward the big gangster. The big gangster briefly returns the look, then stares directly at Dan and Bruiser.
Big gangster asks in English, low and intimidating: {You got the money?}
Transition: Hard cut.

[STAGE 6 | 14–17 seconds | Bruiser gives in]
Camera: Medium-wide two-shot showing Dan, Bruiser and the ground beside the car. Static camera.

Action sequence: Dan silently looks down at Bruiser. Bruiser looks from Dan to the gangsters, hesitates for a beat, then tilts his head in irritated resignation.
Bruiser says in English: {Whatever!}
He swings the sack off his shoulder and drops it at his feet. Its loosely cinched mouth falls open on impact.

End state: The sack rests on the pavement between the two pairs.
Transition: Cut on the sack’s landing.

[STAGE 7 | 17–18 seconds | Cash reveal]
Camera: Tight downward-looking insert of the open sack. Static camera.

Action sequence: The canvas settles, revealing messy bundles of US $100 bills inside. Clearly readable money reveal; the cash stays inside the sack.
Transition: Hard cut.

[STAGE 8 | 18–22 seconds | Opening the trunk]
Camera: Medium-wide side view, tracking laterally along the car from its front quarter toward its rear. Keep the camera outside the trunk, below any sightline into its contents. One continuous shot.

Action sequence: Both gangsters look down at the money and smile. The skinny gangster slides off the hood, walks the short distance to the rear and lifts the trunk lid. He nods toward the open trunk, inviting Dan and Bruiser over.

Dan immediately walks forward towards the trunk. Bruiser briefly freezes, checks the big gangster’s face, then hurries after Dan with tense shoulders and worried LED eyes. The big gangster stays beside the car.

End state: Dan reaches the open trunk first; Bruiser approaches just behind him. The contents remain hidden.
Transition: Hard cut to inside the trunk.

[STAGE 9 | 22–28 seconds | Trunk POV and reactions]
Camera: Fixed wide-angle camera positioned low inside the trunk, looking upward through the opening. The raised lid frames the top of the image. This is an upward-looking trunk POV; the characters look downward into it. Keep every egg below the camera’s field of view.

Initial state: Dan’s head enters the opening first as he leans over the trunk. Bruiser’s shorter head appears a moment later beside him as he reaches the bumper and peers over the edge.
Preserve their height difference.

Action sequence: Dan studies the hidden contents, develops a smug smile and glances sideways down at Bruiser.
Dan says in English: {I told you this is worth it.}
Bruiser’s worried LED eyes expand into large, astonished circles. His body becomes still as fear turns into genuine admiration.
Bruiser asks in English: {Aren’t these banned since the seventies?}

End state: Both remain leaning over the trunk, looking down at the unseen stash.
Transition: Hard cut to their downward view.

[STAGE 10 | 28–30 seconds | Kinder Surprise reveal]
Camera: High-angle close shot looking directly down into the open trunk. A small, controlled push-in.

Action sequence: Reveal the trunk packed with dozens of normal-sized Kinder Surprise eggs matching <<<image_6>>>, messily dumped inside the trunk. Their white-and-red foil and colorful labels are immediately recognizable.

Dan answers from off-camera in English, quietly satisfied: {Yes. Yes, they are.}

End state: Hold on the absurdly large chocolate-egg stash through the end of his reply.
Transition: End.
[VISUAL STYLE]
Grounded cinematic realism with the daylight and gritty textures of <<<image_5>>>. Muted industrial colors, deep natural shadows and the yellow car as the strongest environmental color. Realistic worn metal on Bruiser, convincing skin and fabric textures, restrained film grain. Shoot the entire exchange with the seriousness of a dangerous crime deal. The comedy comes from the final object reveal and the sincere performances.

[CAMERA AND PERFORMANCE]
Clean, purposeful cuts with consistent eyelines and travel direction. Each numbered stage is one uninterrupted shot.

Dan stays relaxed and self-assured: steady pace, loose shoulders, direct eye contact and a small smug smile at the trunk.

Bruiser argues but follows: frequent upward glances at Dan, tense grip on the sack, hesitant steps and expressive blue LED eyes. His speech comes from his robot speaker; his metal faceplate does not become a human mouth.

The gangsters remain intimidating until they see the cash. Keep reactions brief and dialogue conversational but brisk enough for the 30-second sequence.

[AUDIO]
English dialogue, assigned exactly to the specified speakers. Dan sounds casual and confident.
Bruiser sounds nervous and argumentative, with a lightly metallic voice that remains easy to understand. The big gangster speaks in a low, firm voice.

<Quiet distant traffic, footsteps on gritty pavement, subtle robot servos, canvas rustling, a heavy sack landing, trunk latch clicking and trunk hinges moving.>
Keep environmental sound beneath the dialogue.

[MAINTAIN CONSISTENCY]
Preserve all four identities, Bruiser’s short scale, the same yellow car and alley layout, one cash sack, and the hidden trunk contents until the final Kinder Surprise reveal.

→ But I don't like the generated voices. We can take it a step further and create voice consistency

Final prompt with my own voices

Only the audio part changed: Dan matches @Video 1, Bruiser matches @Video 2. Five-second black-screen MP4s.

[GOAL]
A 30-second cinematic crime-comedy scene. Dan and his nervous robot friend Bruiser enter a back alley for a suspicious cash deal with two intimidating gangsters. Play the encounter completely straight, building tension until the final reveal: the yellow car’s trunk is packed with Kinder Surprise eggs.

[REFERENCES]

@Image 1 — Dan’s appearance. Preserve his exact face, tousled light-brown hair, thin mustache, black sweatshirt, black trousers and white sneakers. All panels depict one man.

@Image 2 — Bruiser’s appearance. Preserve his short, stocky robot body, weathered white-gray metal, rust-orange armor, “07” chest marking, antennas and blue LED eyes. Keep his approximately one-meter height relative to the adults. The panels show different views and expressions of the same robot.

@Image 3 — Skinny gangster. Preserve his face, scrawny build, black beanie, oversized gray hoodie, worn graphic T-shirt, baggy jeans, jewelry and sneakers. All panels depict one man.

@Image 4 — Big gangster. Preserve his face, bald head, facial tattoos, muscular build, black bomber jacket, dark clothing, chain and sneakers. All panels depict one man.

@Image 5 — Location and car. Match the gritty graffiti-covered alley, industrial walls, scattered debris, daylight and yellow sports car. Use this reference for the environment and vehicle only; the two gangsters are defined by @Image 3 and @Image 4.

@Image 6 — Kinder Surprise packaging. Match the egg shape, crinkled white-and-red foil and colorful Kinder Surprise branding. The two views show the same product design, repeated across many normal-sized eggs in the trunk.

@Video 1 — Dan’s voice only. Use the voice of Dan speaking in this reference as the consistent voice for every Dan line. Match its vocal identity, timbre, pitch, accent and natural speech rhythm while adapting the delivery to the scene. Do not use this video's visuals, setting, actions, background sounds or original dialogue.
@Video 2 — Bruiser’s voice only. Use the voice of Bruiser speaking in this reference as the consistent voice for every Bruiser line. Match its vocal identity, timbre, pitch, accent, natural speech rhythm and any existing vocal processing while adapting the delivery to the scene. Do not add extra robotic filtering beyond the reference. Do not use this video's visuals, setting, actions, background sounds or original dialogue.

Keep the voice assignments fixed: @Video 1 belongs only to Dan; @Video 2 belongs only to
Bruiser. Neither gangster uses either reference voice.

Do not reproduce reference-sheet panels, studio backgrounds or measurement graphics.

[CONTINUITY]

Exactly four characters: Dan, Bruiser, skinny gangster and big gangster.

Bruiser alone carries one medium-sized, loosely cinched canvas sack over his shoulder. Dan walks beside him empty-handed. The sack contains bundles of US $100 bills and remains on the ground beside the car after Bruiser drops it.

The skinny gangster initially sits on the yellow car’s front hood. The big gangster leans against the adjacent front fender. The trunk is at the rear of that same stationary car.

Keep the Kinder Surprise eggs completely hidden from the audience until the final shot. When the trunk opens, frame its exterior and the characters without exposing the contents.

Dan’s voice remains matched to @Video 1 throughout. Bruiser’s voice remains matched to
@Video 2 throughout, including when his emotion changes from nervousness to irritation to amazement.

[STAGE 1 | 0–5 seconds | Entering and rounding the corner]

Camera: Medium-wide three-quarter frontal tracking shot, retreating with Dan and Bruiser. As they turn the corner, arc continuously beside them into an over-the-shoulder wide view down the alley. No internal cut.
Initial state: Dan walks casually, shoulders loose. Bruiser keeps close beside him, gripping the sack over his shoulder, blue LED eyes flicking toward the surrounding walls.

Action sequence: Before they apss the corner, Bruiser looks up at Dan.

Dialogue language: English. Bruiser, using the voice from @Video 2, asks anxiously: {Are you sure this is the right place?}

Dan gives no answer and keeps walking. They round the corner, revealing the two gangsters waiting at the far end of the alley beside the yellow car: skinny gangster sitting on the hood, big gangster leaning against the fender.

End state: Dan and Bruiser occupy the foreground; the gangsters and yellow car are clearly visible in the distance.

Transition: Hard cut.

[STAGE 2 | 5–6 seconds | Intimidating gangster insert]

Camera: Tight low-angle two-shot from near hood height, looking upward at both gangsters. Static camera.

Action sequence: cinematic unnerving sound effect boom as the skinny gangster stares toward the approaching pair from the hood. Beside him, the big gangster lowers his chin and fixes them with an unblinking glare. Neither speaks or smiles.

Transition: Hard cut.

[STAGE 3 | 6–10 seconds | Nervous approach]

Camera: Medium frontal tracking two-shot, moving backward as Dan and Bruiser approach the car. Frame low enough to keep the short robot fully readable beside Dan.

Action sequence: Bruiser hunches around the sack, glances toward the gangsters, then urgently looks up at Dan.

Bruiser, using the voice from @Video 2, says in English: {I told you this is a bad idea!}
Dan keeps his eyes ahead and answers with effortless calm.

Dan, using the voice from @Video 1, says in English: {Relax, Bruiser. It’s a simple deal.}

End state: They reach the gangsters and slow to a stop.

Transition: Hard cut.

[STAGE 4 | 10–12 seconds | Dan opens the deal]

Camera: Medium close-up of Dan over the big gangster’s shoulder. Static, eye-level framing.

Action sequence: Dan stops beside the front of the car, meets the gangsters’ eyes and gives one sharp upward nod.

Dan, using the voice from @Video 1, asks in English: {You got the good stuff?}

End state: Dan holds his confident expression.

Transition: Reverse-angle cut.

[STAGE 5 | 12–14 seconds | The money question]

Camera: Tight two-shot of the gangsters from Dan’s position, slightly low angle. Static camera.

Action sequence: The skinny gangster turns his eyes toward the big gangster. The big gangster briefly returns the look, then stares directly at Dan and Bruiser.

Big gangster asks in English, in his own low, intimidating voice: {You got the money?}

Transition: Hard cut.

[STAGE 6 | 14–17 seconds | Bruiser gives in]

Camera: Medium-wide two-shot showing Dan, Bruiser and the ground beside the car. Static camera.

Action sequence: Dan silently looks down at Bruiser. Bruiser looks from Dan to the gangsters, hesitates for a beat, then tilts his head in irritated resignation.

Bruiser, using the voice from @Video 2, says in English: {Whatever!}

He swings the sack off his shoulder and drops it at his feet. Its loosely cinched mouth falls open on impact.

End state: The sack rests on the pavement between the two pairs.

Transition: Cut on the sack’s landing.

[STAGE 7 | 17–18 seconds | Cash reveal]
Camera: Tight downward-looking insert of the open sack. Static camera.

Action sequence: The canvas settles, revealing a stack of US $100 bills inside. Clearly readable money reveal; the cash stays inside the sack.

Transition: Hard cut.

[STAGE 8 | 18–22 seconds | Opening the trunk]

Camera: Medium-wide side view, tracking laterally along the car from its front quarter toward its rear. Keep the camera outside the trunk, below any sightline into its contents. One continuous shot.

Action sequence: Both gangsters look down at the money and smile. The skinny gangster slides off the hood, and lifts the trunk lid while the bald gangster slowly walks out of the way. The skinny gangster nods toward the open trunk, inviting Dan and Bruiser over.

Dan immediately walks forward. Bruiser briefly freezes, checks the big gangster’s face, then hurries after Dan with tense shoulders and worried LED eyes. The big gangster stays beside the car.

End state: Dan reaches the open trunk first; Bruiser approaches just behind him. The contents remain hidden.

Transition: Hard cut to inside the trunk.

[STAGE 9 | 22–28 seconds | Trunk POV and reactions]

Camera: Fixed wide-angle camera positioned low inside the trunk, looking upward through the opening. The raised lid frames the top of the image. This is an upward-looking trunk POV; the characters look downward into it. Keep every egg below the camera’s field of view.

Initial state: Dan’s head enters the opening first as he leans over the trunk. Bruiser’s shorter head appears a moment later beside him as he reaches the bumper and peers over the edge.
Preserve their height difference.

Action sequence: Dan studies the hidden contents, develops a smug smile and glances sideways down at Bruiser.

Dan, using the voice from @Video 1, says smugly in English: {I told you this is worth it.}

Bruiser’s worried LED eyes expand into large, astonished circles. His body becomes still as fear turns into genuine admiration.

Bruiser, using the voice from @Video 2, asks in stunned English: {Aren’t these banned since the seventies?}
End state: Both remain leaning over the trunk, looking down at the unseen stash.

Transition: Hard cut to their downward view.

[STAGE 10 | 28–30 seconds | Kinder Surprise reveal]

Camera: High-angle close shot looking directly down into the open trunk. A small, controlled push-in.

Action sequence: Reveal the trunk packed with dozens of normal-sized Kinder Surprise eggs matching @Image 6 stacked in the trunk. Their white-and-red foil and colorful labels are immediately recognizable.

Dan answers from off-camera in English, using the same voice from @Video 1, quietly satisfied:
{Yes. Yes, they are.}

End state: Hold on the absurdly large chocolate-egg stash through the end of his reply.

Transition: End.

[VISUAL STYLE]

Grounded cinematic realism with the daylight and gritty textures of @Image 5. Muted industrial colors, deep natural shadows and the yellow car as the strongest environmental color. Realistic worn metal on Bruiser, convincing skin and fabric textures, restrained film grain. Shoot the entire exchange with the seriousness of a dangerous crime deal. The comedy comes from the final object reveal and the sincere performances.

[CAMERA AND PERFORMANCE]

Clean, purposeful cuts with consistent eyelines and travel direction. Each numbered stage is one uninterrupted shot.

Dan stays relaxed and self-assured: steady pace, loose shoulders, direct eye contact and a small smug smile at the trunk.

Bruiser argues but follows: frequent upward glances at Dan, tense grip on the sack, hesitant steps and expressive blue LED eyes. His speech comes from his robot speaker; his metal faceplate does not become a human mouth.
The gangsters remain intimidating until they see the cash. Keep reactions brief and dialogue conversational but brisk enough for the 30-second sequence.

[AUDIO]

Dialogue language: English.

Dan: Match @Video 1 for every spoken line, including his final off-camera reply. Preserve the reference voice while performing the dialogue casually and confidently.

Bruiser: Match @Video 2 for every spoken line. Preserve the reference voice while moving through nervous questioning, panicked objection, irritated resignation and stunned admiration.
His robot appearance must not override the supplied voice.

Big gangster: A distinct low, firm, intimidating male voice, clearly different from Dan and Bruiser.

Speak only the scripted dialogue. Do not replay dialogue, music or background audio from the voice-reference videos. Keep each line assigned to its specified speaker, with clear turn-taking and synchronized visible speech.

<Quiet distant traffic, footsteps on gritty pavement, subtle robot servos, canvas rustling, a heavy sack landing, trunk latch clicking and trunk hinges moving.>

Keep environmental sound beneath the dialogue.

[MAINTAIN CONSISTENCY]

Preserve all four identities, Dan’s @Video 1 voice, Bruiser’s @Video 2 voice, Bruiser’s short scale, the same yellow car and alley layout, one cash sack, and the hidden trunk contents until the final Kinder Surprise reveal.
09

Motion transfer

Higgsfield Genjutsu

  • Upload the video for motion extraction, confirm the length, add your character sheets, generate
  • That's it. You can add a prompt but you don't need one
  • Film it yourself and put the character on top. Don't steal other people's content

Video reference + character sheets → result

The two character sheets I used

Image 1: ginger character in a brown suitImage 2: Dan character sheet
Copied