五间房 · 第一段 — I Only Wanted a Room, You Want My Life — AI Cinematic Short

五间房 · 第一段 — I Only Wanted a Room, You Want My Life is a 15-second, vertical 9:16 AI film beat set in the interstellar magic-inn world 银河破釜旅馆 (Galaxy Blow-up Inn) in Neta Studio. It is the first segment of the POV comedy short series 五间房 (Five Rooms), shot entirely first-person: the viewer is 晓米, the inn's only full-time employee, following the guests down one corridor. An arrogant noble heir, Ma Xiaofu, storms in demanding the best room — and gets flattened twice: first by a flurry of winged piglets bursting out of the blue door, then by a fireball and a heavy tome flying out of the pink one. Two doors, one very bad night, no room.

Watch it now: 五间房 · 第一段 — I Only Wanted a Room, You Want My Life →

1080×1920 (9:16) · 15s · one continuous first-person walk — rendered with Seedance 2.0 from the world's own locked reference frames, in a warm-brown 3D retro-realistic timber-and-stone cinematic look, with native Chinese dialogue and sound. The corridor never changes: one fixed set, the doors always on the left wall, the camera only ever moving forward.

Download the MP4 →

15 seconds. Photorealistic 3D retro-realistic cinematic render — like a frame from a real-time 3D game engine (Unreal Engine 5) or a photograph of a physical film set: aged real materials (dark timber beams, rough stone, worn plaster, wrought iron, brass, leather), physically accurate warm firelight and window light, deep amber-and-brown filmic colour grading with rich dark brown shadows and glowing warm highlights. Ultra sharp and crisp, deep focus, high clarity in every part of the frame — no blur, no soft focus, no depth-of-field haze, no heavy grain, no glow effects. NOT an illustration, NOT a painting, NOT a cartoon, no drawn lines, no airbrush.

Reference images: the first attached reference image is the material and lighting mood reference — match its warm brown cinematic realism and its real timber, stone and leather materials exactly. The second attached reference image is the inn's interior — keep this same timber-and-stone corridor look with fixed furniture. The third attached reference image is Ma Xiaofu, a pale-violet elf-eared young nobleman with silver-white hair and a deep-purple robe with gold trim — keep his look exactly. The fourth attached reference image is Harley — pale sky-blue elf-eared young man, dark messy hair, round mechanical glasses, black robe and scarf — keep his look. The fifth attached reference image is Hermione — pale sky-blue elf-eared young woman, long dark-brown wavy hair, a metal monocle on her right eye, black robe — keep her look.

FIXED SET — never changes, in any shot: one inn corridor. All doors sit in a row along the LEFT wall of the corridor, in this fixed order — first a BLUE door, then further along a PINK door. The doors stay on the left wall, never swap sides, never move. Lanterns, wall hooks, floorboards, side tables and all furniture stay exactly where they are, in the same positions, all the way through. The camera only ever moves forward down the corridor; it never crosses to the other wall.

Event: the arrogant young nobleman arrives demanding the best room; two guest rooms knock him flat. By the end he has been beaten by two doors and still has no room.

Spatial blocking: the viewer IS 晓米, the inn's young woman on duty, walking behind him — only the corridor, its doors and the rooms are visible, never her face or body; she speaks only as an off-screen voice. Ma Xiaofu walks ahead of the camera along the left-hand doors.

Camera: first-person handheld at eye-level, following forward down the corridor, steady and stable; vertical framing.

Beat timeline:
0 to 4 seconds: warm timber-and-stone inn corridor, lanterns and firelight, sharp deep focus; Ma Xiaofu strides in ahead of the camera, flings his cloak, lifts his chin and sneers "本少爷要住最好的房间！"
4 to 9 seconds: he stops at the BLUE door on the left wall and shoves it open — inside, Harley crouches on the floor dressing a swarm of small colourful winged piglets; the piglets burst out in a flurry straight into Ma Xiaofu's face and blast him off his feet back across the corridor floor; Harley leans out of the doorway, delighted, and says "哇，看得出来，我的小飞猪们很喜欢你哦，马小福！"
9 to 15 seconds: Ma Xiaofu scrambles up, hair mussed, and wrenches open the PINK door, the next one along the same left wall — a fireball and a thick heavy book fly out and explode against his face, his silver hair bursting into a popcorn cloud of smoke; Hermione strolls out slowly with wand and book, pushes her monocle up, and says calmly "哦，马小福，你有兴趣来当我的实验对象吗？"; Ma Xiaofu slams the door shut, presses his back to it, breathing hard, and blurts "走！我们去下一间房！"

Sound: Chinese dialogue plus magical effects and a light comedic score. Wingbeats, piglet squeals, a thud, fireball whoosh, an explosion, fluttering book pages, a door slamming. Each speaker keeps a stable, distinct voice: Ma Xiaofu an arrogant young male voice, Harley a delighted male voice, Hermione a calm female voice, 晓米 a level female off-screen voice. Comedy score is light and upbeat, never drowning the dialogue.

Hard constraints: every character's face, hair, skin, ears, monocle and costume stay identical to their reference images and never change between shots; the corridor, its doors, lanterns and furniture stay fixed and unchanged, all doors on the same left wall; keep the warm brown, 3D retro-realistic wood-and-stone cinematic look throughout; ultra sharp, no blur; no duplicate or extra people; no on-screen text, no subtitles, no captions, no logos, no watermarks; the viewer's face never appears.

Model: Seedance 2.0 — image-to-video from the world's locked reference frames, 15s single first-person walk @ 1080×1920 (9:16), native dialogue and sound.

Neta Studio turns one board into a film. Build the world, lock the look with image models like GPT Image 2, hand the exact reference frames to video models like Seedance 2.0, Seedance 2.0 Fast, Kling, MiniMax H3 and Veo — and every beat comes back with the same face, the same corridor and the same warm-brown grade.

Direct a hard, fixed set. Doors on one wall, a locked force of characters and one forward camera — Neta Studio holds all of it across a 15-second take, with native dialogue and sound. Reject and re-take as much as you need, then share a link that plays anywhere, no download.