Una toma fija.Todos los ángulos.

{
  "video_urls":  ["tu-toma-cortada.mp4"],
  "image_urls":  ["frente.jpg", "perfil.jpg", "plano-medio.jpg"],
  "prompt":      "…",
  "resolution":  "720p",
  "aspect_ratio": "9:16",
  "duration":    "10",
  "generate_audio": false,
  "seed":        1234
}

Keep the person, the room and the timing from @Video1 exactly as they are.
The face and clothing match @Image1 and @Image2.
Re-frame the same moment as a slow push-in from a low angle, ending at
a medium close-up. Handheld drift, very subtle. Natural daylight, no
color grade, no added text. Do not use the background from @Image1.

Same person, same room, same audio timing as @Video1.
Face reference: @Image1, @Image2.

0-4s   wide shot from the side, static.
4-9s   cut to a top-down angle looking down at the desk.
9-15s  cut to a medium close-up, slow push-in, handheld drift.

Keep the subject consistent across all three shots. Natural daylight.

Create a 28.6-second Seedance 2.5 video-to-video transformation using the uploaded footage as the exact performance reference.

Preserve the same room, chair, microphone, clothing, hair, body proportions, facial identity, skin texture, hand gestures, eyeline, lip movement, speech timing and acting. He wears no cap and no hat. He holds a small handheld microphone in his right hand the whole time. The main character's face must remain identical in every frame. Do not beautify, reshape, replace or reinterpret him. Keep every spoken word unchanged with accurate lip sync.

Make it one continuous, dynamic music-video-style take inside the same room. No cuts, scene changes, teleporting, duplicate characters or random objects. Every virtual camera move must connect smoothly with believable perspective, parallax, motion blur and room geometry. Keep the character anchored to his original performance. Do not distort or cover his face. Use polished kinetic supers timed to key phrases, appearing beside or behind him with realistic depth and clean dissolves. Do not subtitle the spoken words — the only on-screen text is the supers listed below, exactly as written.

0-2.65s, "¿Sabías que puedes lograr estas tomas cinematográficas"
Start from the original framing. The daylight and the room lights power off, turning the room dark and creating a dramatic stage silhouette with a thin rim light around his hair and shoulders, subtle haze and a back spotlight. Add a slight floating camera push. Large 3D text "TOMAS DE CINE" briefly forms behind him, partially hidden by his body, then fades. Preserve enough facial detail to maintain identity.

2.65-4.3s, "sin tener que mover tu cámara?"
Accelerate smoothly into a tight face close-up using a physical camera push, not a digital crop. Keep his face sharp and unchanged. "¿SIN MOVER LA CÁMARA?" appears briefly beside him without blocking his face.

4.3-8.2s, "Simplemente tienes que grabarte sentado y completamente quieto con tu tripié"
Normal daylight returns. Pull rapidly backward beyond the original framing and extend the same room naturally, revealing a smartphone on a tripod in the foreground recording him. Correct perspective must show this was a locked static setup. Add "TOMA ESTÁTICA" near the tripod. Do not duplicate the character.

8.2-11.5s, "Pasar esa toma estática a Seedance 2.5"
Lower the virtual camera toward desk height and arc smoothly to one side, ending on a strong side-profile angle. Keep his performance and lip sync exact from the new perspective. "SEEDANCE 2.5" travels through the room with dimensional depth, then dissolves.

11.5-15s, "y describirle ahí el ángulo de la cámara o el movimiento que quieras"
A translucent holographic interface appears beside him. "TOP DOWN" types itself into the interface, glows, then dissolves as the command activates. The camera cranes vertically upward into a high top-down angle while he continues speaking and looks naturally toward the moving camera. Preserve the room layout and avoid ceiling or body warping.

15-18s, "Puede ser algo suave como un gimbal acercándose"
Descend to eye level and perform a clean gimbal move: smooth push toward him, ease backward, then settle. No jerky digital zoom. "GIMBAL SUAVE" passes behind his shoulders.

18-21s, "o algo más agresivo como un dron dando la vuelta alrededor"
Accelerate into one complete 360-degree indoor drone orbit around him. Keep him fixed in the chair and the room consistent. Use believable motion blur and near-object parallax, then land precisely at the original frontal position. No cloning, room mutation or lip-sync interruption.

21-24.6s, "¿O hasta títulos 3D gigantes apareciendo detrás de ti?"
Kinetic 3D words appear behind him in sequence: "3D", "GIGANTE", "DETRÁS DE TI". The letters occupy real space, cast subtle light and shadow, and are correctly hidden where his body overlaps them. They pop outward with impact, then retract and dissolve.

24.6-28.6s, "Te lo juro, nunca vas a tener que contratar un camarógrafo, nunca más."
A cameraman enters naturally from the right side carrying a compact camera rig. Without stopping his speech or changing his eyeline, the main character firmly but playfully pushes and signals him out with his free left hand. The cameraman reacts and exits. "SIN CAMARÓGRAFO" appears behind them, then disappears. Return to the original framing.