Los prompts del video — Ruva
Recursos · Prompts del video

Los prompts del video.

Tres pasos para recrear el efecto con tu propia imagen. Sigue el orden, copia los prompts y obtén el mismo resultado.

01 Personaje
02 Storyboard
03 Video
Paso 01 · ChatGPT Images 2

Guía de personaje

Imagen
Cómo usarlo

Elige a una persona, sube una foto suya a ChatGPT Images 2 y pega el siguiente prompt. ChatGPT te devolverá una hoja de personaje multi-vista con la persona vestida estilo Thriller-era.

A professional character reference sheet of the subject from the uploaded reference photo(s), wearing a Michael Jackson Thriller-era inspired outfit, laid out as a multi-view turnaround on a clean neutral light gray seamless backdrop.

Reference images:
- Images 1 to N: photographs of the subject — use these to extract the subject's exact face, skin tone, hair, facial hair, eyewear, and any other distinguishing features. Ignore the lighting, background, framing, and clothing of the reference photos — only preserve the person's identity.
- Final image (layout reference): multi-view character sheet grid with 7 panels (top row: 4 full-body turnaround views; bottom row: 3 face close-up views; small-caps panel labels below each figure on neutral background).

Subject identity to preserve exactly across all 7 panels:
- Same face, same skin tone, same proportions as in the reference photo(s)
- Same hair (length, texture, color, style) as in the reference photo(s)
- Same facial hair if present (beard, mustache, stubble, or clean-shaven — exactly as in the reference)
- Same glasses or eyewear if present in the reference; if not present, do not add any
- Same age and build as in the reference photo(s)
- Calm neutral expression in every panel
- Do not redesign, idealize, or "improve" the subject's features — render them as they appear in the reference

Wardrobe — Michael Jackson Thriller-era inspired (apply to all 7 panels identically):
- Red leather jacket with black diagonal panel detailing, popped collar, fitted waist, zippered front
- Black slim-fit trousers, cropped slightly above the ankle
- White cotton crew socks, visible
- Black polished leather loafers
- Single sequined white glove on the right hand
- No fedora, no hat

Layout — match the layout reference grid exactly:

Top row (four full-body panels, full figure head to toe, equal spacing):
- Panel 1: "FRONT" — facing camera, arms relaxed at sides, gloved hand visible
- Panel 2: "SIDE PROFILE" — left-facing 90° profile, arms relaxed
- Panel 3: "BACK" — back view, arms at sides, jacket back paneling visible
- Panel 4: "BACK 3/4" — three-quarter back view from the right side

Bottom row (three face close-up panels, head and shoulders only):
- Panel 5: "FACE FRONT" — direct front portrait, neutral expression
- Panel 6: "FACE PROFILE" — left-facing 90° face profile
- Panel 7: "FACE 3/4" — three-quarter face turn to camera-left

Panel labels below each figure in small-caps sans serif type, verbatim: "FRONT", "SIDE PROFILE", "BACK", "BACK 3/4", "FACE FRONT", "FACE PROFILE", "FACE 3/4". Render text verbatim — no extra characters, no substitutions.

Lighting and rendering:
- Flat even studio lighting, soft diffuse, no dramatic shadow
- Neutral light gray seamless backdrop, consistent across all panels
- Photorealistic, real skin texture, natural detail, sharp focus
- Consistent lighting, color temperature, and scale across all 7 panels
- Do NOT carry over the lighting, color cast, or environment from the reference photos — render with neutral flat studio light regardless of how the references were shot

Use case: character reference sheet for content production.

Constraints:
- Preserve the subject's exact face, hair, facial hair, and eyewear from the reference photo(s) across all 7 panels — do not redesign the character, do not change ethnicity, age, or features
- No facial-feature drift between panels, no identity drift
- Do not attempt to replicate Michael Jackson's face or likeness — only the wardrobe is inspired by his Thriller era; the subject retains their own identity from the reference photo(s)
- No watermark, no extra text, no background props, no extra accessories beyond the wardrobe specified
- Identical wardrobe styling in every panel — no variations between views
- If the reference shows the subject without glasses, do not add glasses; if with glasses, keep the same glasses style
Desliza para ver el prompt completo
Paso 02
Paso 02 · ChatGPT Images 2

Storyboard 16 paneles

Imagen
Cómo usarlo

Sube la imagen que te devolvió ChatGPT en el paso anterior (la guía de personaje) y pega este prompt. ChatGPT generará un storyboard fotorrealista de 4×4 paneles con los 16 movimientos del baile.

A 16-panel photorealistic stage performance storyboard, arranged as a clean 4x4 grid on a thin dark charcoal divider background. Each panel is a high-end concert photography frame of the same performer captured mid-movement on a professional theatrical stage with cinematic lighting.

Image 1: character + wardrobe anchor — the uploaded character reference sheet. Preserve the subject's exact face, skin tone, hair, facial hair, eyewear, and the COMPLETE wardrobe (every garment, every accessory, every detail exactly as shown). This is the absolute identity and outfit lock for all 16 panels — do not alter, add, or remove any element.

Image 2: layout reference — use ONLY the grid structure (4x4 panels, large stylized numeral in the top-left corner of each panel, small-caps move-name label below each panel, header banner across the top). DO NOT replicate the visual style of Image 2 — each panel must be a photorealistic concert photograph, not an illustration.

Header banner across the top of the entire composition, in bold all-caps sans serif type (render verbatim): "POP STAGE ROUTINE — 16 COUNTS — 10 SECONDS — LIVE PERFORMANCE FLOW"

Each panel must render as a photorealistic stage performance photograph with these consistent visual properties:
- Professional theatrical stage with a dark backdrop (no audience visible in detail, only implied darkness or silhouetted heads at the very bottom edge in a few panels)
- Cinematic stage lighting: warm white key spotlight from above, cool blue rim light from behind, occasional magenta or amber accent lights
- Visible volumetric haze and light beams cutting through the air
- Polished dark stage floor with subtle reflections of the performer
- Shot on a 35mm cinema camera feel, shallow depth of field, real skin texture, subtle film grain
- Sharp focus on the performer, motion blur only on fast-moving limbs where indicated
- Consistent color temperature and lighting setup across all 16 panels

Each panel — large stylized numeral (1–16) in the top-left corner in bright white or red sans serif, small white motion-direction arrow within the frame (subtle, not illustrated), and a small-caps label centered below the panel in clean sans serif (render all labels verbatim).

Row 1:
1. "TOE STAND" — performer balanced on the tips of both feet, body upright, arms extended laterally at shoulder height, key light from above, dramatic shadow on stage floor
2. "FORWARD LEAN" — performer leaning forward at an extreme angle, feet flat anchored to stage, arms slightly out, side-lit with strong rim light
3. "BACKWARD GLIDE" — performer captured mid-step sliding backwards, one foot flat on floor, the other heel raised toes pointing down, slight motion blur at feet
4. "SHARP SPIN" — performer mid-360° rotation on one foot, the other leg tucked up, arms close to chest, slight rotational motion blur at the wardrobe edges

Row 2:
5. "SIDE GLIDE LEFT" — performer sliding sideways to camera-left, feet apart, weight transferring left, right arm bent up at the elbow, subtle floor reflection
6. "SIDE GLIDE RIGHT" — mirror of panel 5, sliding to camera-right
7. "HEEL POP" — performer on the balls of both feet with heels lifted, arms crossed at chest
8. "KNEE DROP" — performer captured mid-descent landing on one knee, opposite leg extended back, right arm raised dramatically toward the spotlight

Row 3:
9. "HIGH SIDE KICK" — performer mid-air with one leg kicked out high to the side at hip-height or higher, opposite arm raised, freeze-frame sharpness, dust kicked up at takeoff foot
10. "SHOULDER POP" — performer with right shoulder thrust forward, left arm bent at 90°, head tilted, sharp side lighting carving the wardrobe silhouette
11. "CROUCH SLIDE" — performer in low crouch sliding sideways across stage, one hand near the floor for balance, motion blur at trailing foot
12. "ARM REACH" — performer standing tall, right hand extended high above the head with fingers spread toward a spotlight, left hand low at the hip

Row 4:
13. "CROSS STEP" — performer mid-step with legs crossed at the knees, arms swung in opposite directions, slight motion blur at the legs
14. "SHARP FREEZE" — performer locked in an angular freeze pose, one arm across the body, head turned sharply toward camera, hard rim light defining the silhouette
15. "ROBOT LOCK" — performer in a mechanical right-angle pose with both arms bent at 90°, knees slightly bent, completely still under hard top-down spotlight
16. "FINAL POSE" — performer balanced on tiptoes with right arm extended diagonally up toward the lighting rig, head tilted slightly, a single bright back-spotlight creating a halo silhouette around the figure

Across all 16 panels — preserve identically:
- Same face, hair, facial hair, eyewear, and skin tone exactly as in Image 1
- Same complete wardrobe from Image 1 in every panel — every garment, every accessory, no item missing, no item added, no styling variation
- Photorealistic concert photography rendering, not illustration
- Consistent theatrical stage environment (dark backdrop, haze, beam lighting)
- Consistent color grading across the grid

Render all text (header and panel labels) verbatim — no extra characters, no substitutions, no extra words.

quality: high.

Use case: dance choreography reference storyboard for content production.

Constraints:
- Preserve the subject's exact face from Image 1 in every panel — no facial-feature drift, no identity drift, no aging
- Each panel must be a photorealistic photograph — NOT a drawing, NOT an illustration, NOT a render
- Do not replicate the visual style of Image 2 — only use its grid layout, numbering, and labeling structure
- Identical wardrobe in every panel as shown in Image 1 — no variations, no missing pieces, no added accessories
- No audience faces, no recognizable real venue, no logos, no watermarks
- Generic original choreography — do not reference, replicate, or evoke any specific real public figure or copyrighted performance
- No duplicate text, no extra labels, no extra panels — exactly 16 numbered panels
Desliza para ver el prompt completo
Paso 03
Paso 03 · Seedance 2.0

Animación 10 segundos

Video
Cómo usarlo

Entra a cualquier plataforma que use Seedance 2.0, sube las dos imágenes anteriores (guía de personaje + storyboard) y pega este prompt. Recibirás un video de 10 segundos con tu personaje haciendo el baile completo.

[REFERENCES]
@Image2 = character identity lock (face, hair, eyewear, facial hair, complete wardrobe — use this as the ONLY identity and outfit source)
@Image1 = choreography storyboard (follow the 16 numbered panels in exact sequence)

[SUBJECT]
The exact person from @Image2. Preserve face, skin tone, hair, facial hair, eyewear, and the COMPLETE wardrobe (every garment and accessory) exactly as shown in @Image2. Identity, facial features, and wardrobe must remain pixel-consistent with @Image2 across every frame — do not add, remove, or alter any item.

[SCENE]
Professional theater stage. Polished dark wood floor with subtle specular reflections. Deep matte-black backdrop. Atmospheric haze and light volumetric smoke catching the beams. Cool blue and white spotlights from upstage left and right creating strong rim light on the subject, with a warm key light from front-camera-right. No visible audience, only faint silhouetted heads at the very bottom edge.

[ACTION — 10 seconds, 16 counts at ~0.625s per count, follow @Image1 panels in order]
0.00–0.62s (Count 1, panel 1): TOE STAND — subject rises onto the balls of both feet, arms extended horizontally in a wide T-pose, chest forward, chin up.
0.62–1.25s (Count 2, panel 2): FORWARD LEAN — sharp anti-gravity-style forward body tilt, arms angled back, feet anchored, weight forward.
1.25–1.87s (Count 3, panel 3): BACKWARD GLIDE — smooth backslide across the floor moving camera-left, feet appearing to float, upper body upright and still.
1.87–2.50s (Count 4, panel 4): SHARP SPIN — fast 360° spin on the ball of one foot, wardrobe edges flare outward, motion blur on rotation, lands cleanly.
2.50–3.12s (Count 5, panel 5): SIDE GLIDE LEFT — low lunge stance gliding leftward, lead arm pointed down-left, right hand raised at chest.
3.12–3.75s (Count 6, panel 6): SIDE GLIDE RIGHT — mirrored low lunge gliding rightward, opposite arm leading, fists loosely clenched.
3.75–4.37s (Count 7, panel 7): HEEL POP — feet together, both heels lift and snap down in a rhythmic pop, arms folded across chest.
4.37–5.00s (Count 8, panel 8): KNEE DROP — drops onto right knee in a controlled landing, left arm thrust diagonally upward, head tilted up toward the light.
5.00–5.62s (Count 9, panel 9): HIGH SIDE KICK — rises and snaps a high lateral kick with the right leg, arms counterbalanced, torso braced.
5.62–6.25s (Count 10, panel 10): SHOULDER POP — sharp isolated shoulder pop and head turn to camera-right, right hand near collar.
6.25–6.87s (Count 11, panel 11): CROUCH SLIDE — drops into a deep crouch and slides camera-right, right hand brushing the floor.
6.87–7.50s (Count 12, panel 12): ARM REACH — rises and reaches the right arm fully extended toward the upper light, gaze following the hand.
7.50–8.12s (Count 13, panel 13): CROSS STEP — pivoting cross-step footwork, body rotating slightly, dust scuff at the feet.
8.12–8.75s (Count 14, panel 14): SHARP FREEZE — abrupt freeze in a closed defensive stance, fist clenched, jaw set, body completely still for one beat.
8.75–9.37s (Count 15, panel 15): ROBOT LOCK — staccato robot-style isolation, segmented arm and shoulder locks with hard stops.
9.37–10.00s (Count 16, panel 16): FINAL POSE — explosive final pose, right arm fully extended toward the upper-right spotlight, body angled, white-hot lens flare on the extended hand. Hold the pose to the cut.

[CAMERA]
0.0–2.5s: locked wide full-body shot, slight slow push-in.
2.5–5.0s: smooth lateral tracking shot following the side glides and kick, medium framing (knees-up).
5.0–7.5s: dynamic interplay — quick medium close-up on the shoulder pop, low-angle wide on the crouch slide, tilt-up on the arm reach.
7.5–10.0s: pull back to a wide stage shot, then a controlled push-in that lands tight on the final pose at 9.4s.
Anamorphic 2.39:1 feel inside a 16:9 frame, shallow but stable focus, no handheld shake.

[STYLE / LOOK]
Live concert cinematography. Crushed blacks, saturated wardrobe colors as shown in @Image2, cool teal-cyan shadows, warm key highlights, fine film grain, subtle anamorphic horizontal lens flares on the spotlights. Natural motion blur on spins and kicks. Photoreal, not stylized.

[AUDIO]
BGM: percussive electro-funk-pop instrumental at ~117 BPM, punchy kick on every count, syncopated bass, bright synth stabs, no vocals.
SFX: crisp shoe taps on wood, soft whoosh on each spin, fabric snap on wardrobe flares, sharp snare hit synced exactly to the SHARP FREEZE (8.12s) and the FINAL POSE (9.37s).
Ambience: faint distant crowd presence, no cheering, low room tone.
Mix: stereo, music slightly ducked under footwork SFX, headroom for later voiceover.

[HARD CONSTRAINTS]
- Identity must match @Image2 exactly in every frame: facial features, hair, facial hair, eyewear, skin tone.
- Wardrobe must remain complete and consistent across all 16 counts exactly as shown in @Image2 — every garment, every accessory, no item missing, no item added, no styling variation.
- Choreography must follow the 16 numbered panels of @Image1 in exact sequential order.
- No artist names, no brand logos, no on-screen text, no watermarks, no subtitles.
- No visible audience faces, no extra performers in frame.
- No facial distortion, no extra fingers, no melting hands, no wardrobe color drift.
- Maintain continuous motion between counts — no freeze frames except the explicit SHARP FREEZE (count 14) and the held FINAL POSE (count 16).
- Stage lighting and floor reflections must stay consistent throughout.
- Camera must not shake or wobble.
Desliza para ver el prompt completo
Próxima edición · Cupos limitados

¿Quieres dominar la IA y diseñar 10× más rápido?

Muy pronto abriremos las puertas a Master Lab IA, nuestra academia especializada. Cupos limitados — entra a la lista de espera para acceso prioritario.

Las puertas cierran tras la apertura
Entrar a la lista de espera
Acceso prioritario · Sin compromiso
Ruva Academy · 2026