Set
A soundstage that never pretends to be anywhere else. A hand-painted city skyline hangs upstage with warm window lights, a zebra crossing is painted across the floor, and the lighting stands and a red postbox stay visible at the edges of frame.
One routine, any performer
Most AI video prompts fail in the places you did not think to specify, and that is most places. A template inverts the problem: the stage, the blocking, the key light, the costume contrast and the lens are already written and already work. You change one thing — who is leading — and it comes out recognisably the same number with somebody else at the front. That is what makes a format spread: every version is a punchline against a reference the audience already has.
Sube una foto para cada papel — o describe al personaje e Image-2 los pinta primero (15 créditos por papel pintado, requiere Premium). El protagonista es obligatorio; los bailarines de fondo y el resto caen en la multitud de stock si se dejan vacíos. La banda sonora, la puesta en escena y la cámara permanecen exactamente como en el original — solo cambian las personas.
Recreate the reference video shot for shot — the same dark soundstage, the red postbox, the wooden slatted bench, the zebra crossing painted across the floor and the hand-painted city skyline with warm window lights, the same camera work, blocking, lighting, warm amber grade and edit rhythm. Only the cast changes. The seated lead who gets their cigarette lit and then strides to the centre of the crossing to lead the routine is a tired accountant in a rumpled grey suit, briefcase still in hand. The group crowding the bench with raised lighters is a group of women in sequinned evening dresses. The dancers circling the lead in the wide are an ensemble of dancers in muted grey suits. Everything else stays faithful to the reference video.
The staging in the reference clip — one lead dead centre on a painted crossing, an ensemble fanned symmetrically behind, a hand-painted skyline upstage — comes from a 蔡徐坤 (Cai Xukun) music video. Chinese internet users re-cut it to jumpstyle and other tracks until it became one of the most remixed video formats on the Chinese internet, usually filed under 只因你太美 or the ikun meme.
It spread for a structural reason worth naming, because it is the reason this template exists. The shot is a locked wide with a symmetrical ensemble and a single lead in the only light-coloured costume. That composition survives almost any substitution: change who is at the front and the frame still reads, still balances, and still tells you instantly who to watch. Formats that survive substitution are the ones that get remixed.
This template reproduces that staging grammar — the locked camera, the symmetry, the tonal separation, the skyline and crossing — and nothing else. What you generate is an original character you supply, on a set built to the same spec. It is not a face swap and it does not reconstruct anyone: the reference clip is here as a visual brief, the way a director shows a lookbook.
Estas son las cláusulas que no tienes que acertar, y la razón por la que cada versión sigue pareciendo la misma película. Cada una está en el prompt de arriba, y cada una es editable una vez que estás en el espacio de trabajo.
A soundstage that never pretends to be anywhere else. A hand-painted city skyline hangs upstage with warm window lights, a zebra crossing is painted across the floor, and the lighting stands and a red postbox stay visible at the edges of frame.
The bench beat plays in a tight medium, the dance in a locked wide at chest height holding the whole ensemble. The performers move; the frame does not.
Beat one: the lead sits on the bench while the lighters crowd in shoulder to shoulder, flames up. Beat two: the lead strides to dead centre on the crossing and the dancers circle them, everyone square to camera. Symmetry is what makes the lead readable at a glance.
The lead wears the only light-coloured costume in frame; the ensemble is in muted greys, browns and black. That single contrast is what keeps the lead legible in a wide shot — swap it and the shot stops working.
Hard top key on the lead, warm practical spill from the painted skyline, deep falloff to black at the edges of the stage.
Warm amber highlights against crushed blacks, mild film grain. Contrast pushed hard enough to survive a phone screen at thumbnail size.
Nada aquí todavía
Nadie ha publicado una versión de esta plantilla aún, lo que significa que la primera que se publique es la que todos los demás verán en la parte superior de esta página.
Because a wide shot with nine people in it needs one obvious focal point, and colour does that faster than position. Put the lead in the same tone as the ensemble and the eye has nowhere to land — the routine stops reading as a solo with backup and becomes a crowd.
Camera movement is where generated video falls apart — it drifts, warps the set, and pulls attention away from the only thing you changed. A locked frame makes the subject the entire event, which is exactly what a swap template needs.
No. These templates are built for original and fictional characters. Generating a real, identifiable person performing something they never did is a likeness problem however funny the result is, and it is not what this template is for. Invent someone instead — the mismatch is usually funnier anyway.
Yes. The whole prompt is handed to the video workspace and every clause is editable there. Start from the template, then change one clause at a time so you can actually tell what moved the result.
The template and the prompt are free. You only spend credits when you generate, at the normal rate for whichever model you pick.
Yes. What you generate is yours to use, including commercially. The commercial use terms have the details.