4 - Generate the Video
Using Seedance 2.5 inside @VibloAI, upload all 3 reference images + This Prompt:
SCENE & CHARACTER CONSISTENCY
One continuous, photorealistic 7-second handheld camcorder recording inside an old Mediterranean coastal villa in southern Italy, overlooking the sea during a warm late afternoon.
THE WOMAN: The young brunette woman from the attached female character sheet. Match her exact facial features, natural tan skin, hazel-brown eyes, loosely tied brown hair with flyaway strands, feminine slender build, white spaghetti-strap ruched camisole, loose white linen pants, and minimal jewelry. Preserve her appearance throughout the entire video.
THE MAN: The ginger-haired man from the attached male character sheet. Match his exact face, curly ginger hair, athletic build, black fitted long-sleeve shirt, white pants, and brown loafers. He is completely absent from the shot until his final dialogue. His appearance must remain identical to his character sheet.
The location features aged off-white plaster walls, reddish terracotta floor tiles, and tall white wooden French doors with glass panes, already open onto a small stone balcony. Outside the doors is a beautiful Mediterranean coastal terrace with a small round cafรฉ table, two woven rattan chairs, a dark wrought-iron railing, distant cypress trees, and an uninterrupted view of the blue sea and hazy coastline.
The camera starts inside the room, looking diagonally toward the open French doors. The woman is close to the camera on the right, turned slightly away with her head looking back directly into the lens. The bright sea and sunlit terrace fill the left and center background.
Warm afternoon sun streams through the open doorway. A harsh built-in digital camera flash illuminates the woman's face and body, creating realistic bright highlights on her skin, while the interior behind her remains comparatively dark. The exterior is slightly overexposed. Lighting must look physically natural, with realistic flash falloff, imperfect exposure, slight grain, natural shadows, and the soft texture of early-2000s digital camera footage.
CAMERA & MOVEMENT
Filmed entirely from the perspective of a person holding a small early-2000s consumer camcorder in one hand.
The camera is handheld throughout, with constant tiny natural shakes, subtle horizon tilts, imperfect walking movement, slight focus breathing, and quick reactive reframing. It should never appear mounted on a tripod or stabilized gimbal.
As the woman walks toward and through the open French doors, the camera follows her naturally, moving a few steps forward and slightly right, gradually revealing more of the balcony, ocean, and bright outdoor architecture.
She remains the main focus throughout her dialogue. The camera operator reacts naturally to her movements, keeping her face visible without making the framing look overly composed.
Near the very end, the camera makes one quick, spontaneous handheld pan toward the right as the man suddenly steps into the foreground. There is a tiny moment of natural motion blur and autofocus adjustment before his face comes into clear view.
No cuts, transitions, slow motion, floating camera movements, cinematic dolly shots, or artificial zoom transitions.
PERFORMANCE & EXACT DIALOGUE
0:00โ0:03.3 โ WOMAN'S FIRST LINE
The woman looks directly at the camera over her shoulder with a confident, naturally flirtatious expression. She begins speaking immediately, with no silent introduction.
She says, clearly and completely:
"No one is telling you how to make your AI look like this."
Her delivery is quick, smooth, and conversational, with every word pronounced clearly. As she speaks, she gradually turns her torso toward the camera while taking small natural steps toward the open balcony doors.
Her body language is relaxed, subtly seductive, and effortless. She uses small, feminine hand movements while speaking, keeping her shoulders relaxed and maintaining playful eye contact.
On the exact words "look like this," she raises one hand into an elegant, open-palmed gesture, subtly presenting herself and the beautiful surroundings.
She finishes the entire sentence. Do not cut off the word "this."
0:03.3โ0:03.5 โ NATURAL PAUSE
She takes one small additional step toward the terrace and briefly glances toward the ocean. A very short conversational pause separates her first sentence from the next.
The camera continues moving forward with her.
0:03.5โ0:05.9 โ WOMAN'S SECOND LINE
She turns her face back toward the camera and says, clearly and completely:
"The skin, light, that old camera look?"
Her hand movements must follow the spoken words precisely:
On "The skin," she gracefully raises her hand toward her cheek, lightly indicating her face and skin without covering her features.
On "light," she extends one arm outward toward the ocean and the sunlight pouring through the open doors, briefly presenting the naturally illuminated environment.
On "that old camera look?" she turns back toward the lens, gives a small playful smile, and makes a subtle open-handed gesture toward the camera itself.
The gestures must blend together smoothly as a single natural performance, never appearing robotic, exaggerated, or disconnected from the dialogue.
She keeps walking at an ordinary, unhurried pace, arriving just outside or at the threshold of the open balcony doors as her final word ends.
Her expressions should be softly flirtatious and expressive, with subtle eyebrow movements, natural blinking, delicate smiles, and realistic facial motion.
0:05.9โ0:07.0 โ MAN'S SUDDEN ENTRANCE
Immediately after the woman finishes the word "look?", the man suddenly steps into frame from the camera's RIGHT side, very close to the lens.
This is the first moment he is visible in the entire video.
The camera operator quickly swings the camera toward him in a reactive handheld movement. He confidently leans into the shot, filling much of the foreground while the woman and Mediterranean balcony remain partially visible behind him.
He looks directly into the lens and says, loudly, confidently, and clearly:
"But I will."
His entrance should feel sudden, charming, and slightly cheeky, like someone spontaneously interrupting a friend's recording.
He must finish the entire sentence, including the final word "will." Hold the frame just long enough for his last word to be fully heard before the clip ends.
VOICE & ACCENT
WOMAN'S VOICE: A beautiful, youthful adult woman's voice with a soft, warm, airy, slightly breathy tone. Her voice is smooth, melodic, charming, and lightly flirtatious, with a subtle natural Italian-influenced English accent. Her pitch is medium-high, feminine and pleasant, with a lightly husky texture underneath the softness.
She speaks English fluently, with a gentle Mediterranean rhythm and relaxed musical intonation. Her voice has personality and natural vocal variation, not artificial or overly polished. She sounds like a real young woman speaking casually while walking with a friend.
Her delivery is slightly fast to fit the short clip, but conversational, expressive, and completely intelligible. No exaggerated pauses, drawn-out vowels, or slow, dramatic speech.
MAN'S VOICE: A masculine, rich, lower-pitched voice with a subtle warm Italian-influenced English accent, similar in regional character to the woman's. His voice is deep, attractive, confident, and slightly raspy, with excellent vocal clarity and a stronger projection than hers.
He delivers "But I will" with a confident, playful emphasis on "I", and a decisive finish on "will." His words are brisk and powerful, without shouting unnaturally.
Both voices must sound convincingly human, with real breath, natural resonance, and accurate lip synchronization.
AUDIO โ STRICTLY NO MUSIC
Only realistic environmental and human sounds:
Very soft Mediterranean sea breeze drifting through the balcony.
Subtle footsteps on the terracotta interior tiles and outdoor stone floor.
Faint clothing movement as the woman walks and gestures.
Light natural environmental ambience from the coastline.
Slight movement noise from the handheld camera operator.
Both characters' voices must remain clear and prominent.
ABSOLUTELY NO MUSIC. No soundtrack, background score, singing, humming, melodic ambience, electronic sound effects, artificial whooshes, or cinematic risers.
CRITICAL DIALOGUE & GENERATION RULES
ALL THREE SPOKEN LINES MUST BE DELIVERED, IN ORDER, COMPLETELY AND VERBATIM.
Woman, line 1:
"No one is telling you how to make your AI look like this."
Woman, line 2:
"The skin, light, that old camera look?"
Man, line 3:
"But I will."
Do not omit, shorten, paraphrase, reorder, substitute, or repeat a single word.
Do not insert extra words such as "the" before "light." Do not change "that old camera look" to another phrase.
The woman must deliver BOTH of her complete lines before the man enters. The man must deliver ONLY his assigned line. Their dialogue must never overlap.
No trailing silence, long pauses, stuttering, repeated phrases, or missing final words. Prioritize complete dialogue over unnecessary movement or visual flourishes.
Maintain precise mouth movement synchronized to the correct character's spoken words.
The woman stays recognizable as the same person throughout; her face, hair, outfit, body proportions, and skin tone must not change.
The man's face, clothing, and proportions must match his character sheet exactly. He must never appear early, including as a background person, reflection, or duplicate.
The Mediterranean villa, doorway, balcony furniture, ocean horizon, and surrounding architecture must remain spatially consistent throughout the continuous camera movement.
No captions, subtitles, visible text, logos, watermarks, additional people, artificial facial expressions, overly exaggerated gestures, CGI-looking skin, studio lighting, beauty filters, or cinematic color grading.
FINAL RESULT: A convincing spontaneous seven-second early-2000s handheld home-video recording, with beautiful natural lighting, subtly flirtatious acting, precisely synchronized gestures, complete dialogue from both characters, a sudden confident male entrance, and absolutely no background music.
crazy...that i made this with Seedream 5.5 and Seedance 2.5
nobody is teaching you the actual workflow behind these AI realistic videos..
BUT I WILL, full breakdown, every prompt included๐
3 - Sexy Version of Me
Take a Picture of yourself
And upload it along with this prompt:
Transforming the person in the image into an exceptionally attractive, high-fashion model version of themselves. Subtly enhance facial harmony, symmetry, jawline definition, chin projection, cheekbone structure, midface proportions, and overall attractiveness.
Layout: three evenly spaced panels against a plain medium-gray studio background.
Left panel: Chest-up close-up portrait
Center panel: Full-body front-facing view
Right panel: Full-body back-facing view.
Realism: Extreme photorealism with visible skin pores, subtle imperfections, realistic hair strands, natural facial details, authentic fabric textures, and physically accurate lighting and shadows. Premium fashion campaign photography quality, with striking yet believable beauty.
Step 3: Make the video using seedance 2.5
(it's expensive, but definitely worth it)
upload the first frame as image 1 and the character sheet as image 2
settings: 1080p, 10 sec, 4:3
-------
SCENE: The same beautiful young woman from the attached character reference, filming herself on the Mediterranean balcony from the location reference. Preserve her exact face, long golden-brown hair, white strapless outfit, layered gold necklaces, skin texture and proportions. 4:3 horizontal framing, authentic early-2000s digital camera vacation footage. Bright natural coastal sunlight, slightly overexposed highlights, subtle digital softness, realistic skin texture, handheld imperfections, and natural colors. She is holding the camera herself in one hand like a real selfie video. The framing changes only because she physically moves her arm closer or farther away. No digital zoom. No cuts. The background stays alive the whole time: ocean water shimmering and moving, banana leaves swaying, coastal breeze moving loose hair strands, distant birds occasionally passing. Keep it feeling real and spontaneous.
SHOT: One continuous 9โ10 second selfie video. She begins speaking immediately at frame one with no silence. She must say this script exactly, once only, in this exact order: "No one is telling you how to make AI characters look this real... THE skin, hair, the way they talk and move. So we decided to expose how they do it. Save this." Critical script rules: - She must clearly say โTHE skinโ - Do not skip โTHEโ - Do not repeat any words - Do not add extra pauses - Do not improvise or change the ending - The whole script should flow quickly and naturally, fitting within about 9โ10 seconds
ACTION SEQUENCE SYNCED TO THE WORDS: Start at normal armโs-length selfie framing, medium close-up. She looks into the lens with a confident, attractive, natural expression. Keep the strong intro feel from the previous version. She subtly shifts her shoulders and posture and has natural blinking and small face movement. On โTHE skin,โ she quickly brings the phone physically closer to her face with her arm, creating a close handheld view of her cheek, eye, nose and lips. This is caused by her arm moving the phone closer, not by camera zoom. The movement can briefly soften focus, then recover. Show real pores, peach fuzz, slight redness, tiny imperfections, natural lip texture and realistic skin detail. Immediately after saying โTHE skin,โ she pulls the phone back out right away to normal selfie framing before she says โhair.โ She must not stay close during โhair.โ On โhair,โ the camera is now back out at a normal framing so her hair is visible. With her free hand, she lightly grabs or touches a section of her hair and gently pulls it forward while continuing to speak. The hair should move naturally in the breeze. On โthe way they talk and move,โ she returns fully to the comfortable armโs-length selfie framing and starts fluidly moving her body in a cute-girl pose while still talking. She brings the back of her free hand near her chin, squints slightly in a cute/flirty way, tilts her head a little, and shifts her shoulders/body softly and naturally. This movement should feel feminine, playful, and fluid, not stiff or exaggerated. On โSo we decided to expose how they do it,โ keep the framing normal and let her continue with smooth natural body language and confident eye contact. On โSave this,โ she gives a subtle confident finishing expression directly to camera. No big extra gesture, no awkward pause, no repeated phrase. Throughout the whole shot, keep her body naturally alive: - natural blinking - realistic lip sync - soft posture changes - small shoulder movement - breathing - hair moving in the wind - no freezing - no robotic motion - no overacting
VOICE / AUDIO: Keep the same voice quality that worked well before: a distinctive, hyper-realistic young Southern California womanโs voice. Lower-mid pitch, warm, smooth, lightly husky, subtly breathy, attractive and confident, not bubbly or overly cute. Slightly sultry but still natural. Realistic conversational rhythm, fast enough to fit the full script into 9โ10 seconds, with minimal pauses and no drawn-out delivery. The delivery should feel quick, fluid, and effortless, like she is naturally talking to her phone camera. No generic AI narration, no robotic pacing, no exaggerated Valley Girl accent, no stiff enunciation. Natural outdoor ambience only: soft breeze, distant birds, leaves rustling. No music, no captions, no subtitles, no text overlays. One continuous 4:3 video, around 9โ10 seconds total, with no cuts and no digital zoom. Important timing: the close framing happens only during the words โTHE skin,โ then immediately returns to normal framing for โhair.โ โHairโ must be said while her hair is visible and she is touching it. โThe way they talk and moveโ must happen at full armโs-length framing while she does a cute pose with the back of her hand near her chin.
Step 2: Create a Character Sheet to lock the consistency.
Most people mess up their character sheets...
You need 3 Shots:
- 1 Close Up
- 2 Full Body Front
- 3 Full Body back
Use This Prompt + the first image inside GPT 2.6 Sunburst:
Create a photorealistic 3-panel character sheet of the same woman from the reference image, with strong identity consistency across all three panels. The layout should be a clean horizontal 3-panel sheet on a plain neutral grey studio background. Make the panels evenly spaced and clearly separated, with the character shown consistently in the same outfit, proportions, facial structure, skin tone, and hairstyle.
Panel layout:
Close-up portrait โ facing directly forward, head upright, not tilted, shoulders visible, neutral expression, looking straight into camera.
Full body front view โ standing straight, facing forward, full body visible from head to toe.
Full body back view โ standing straight, facing away from camera, full body visible from head to toe.
Character details:
A beautiful young woman with long light golden-brown hair, soft middle part, smooth straight texture with a few natural flyaways, luminous warm-toned skin, soft blush, glossy full pink lips, delicate nose, almond-shaped eyes, subtle eyeliner, defined cheekbones, elegant feminine facial structure, and a slim toned figure. She should resemble the reference image closely.
She is wearing a white strapless bandeau-style top and a clean, minimal, elegant full outfit that feels believable with the reference โ keep it simple, fitted, and cohesive, with a refined summer-luxury aesthetic. Add delicate layered gold necklaces, including small cross pendants, and keep jewelry consistent across all panels.
Style / quality:
Emphasize ultra-realistic details and human imperfections:
realistic skin texture and pores
subtle peach fuzz
tiny asymmetries in the face
natural lip texture
soft under-eye detail
realistic collarbones, shoulders, arms, and hands
a few stray hairs and baby hairs
realistic fabric folds and tension
accurate anatomy and proportions
Important consistency instructions:
Maintain the exact same character identity in all 3 panels. Keep:
same face
same hair color and length
same body proportions
same outfit
same jewelry
same skin tone
same realism level
The sheet should feel like a professional character turnaround/reference sheet, with clean studio lighting, minimal shadows, no dramatic color grading, no stylization, no extra props, no environmental background elements, and no text labels.
paying influencers $1k for 200 views?
look at this. $0.20 for a 7-sec video.
a yapping AI influencer I made with Wan 3.0 Prime Using 1 reference image I found on Pinterest.
comment "workflow" and I'll send it over.