TOKYO - PAPER CUTS (PROMPTS)
Copy a prompt, paste it into your tool, change the city.
1. The four stills
Any GPT Image 2 tool. Attach one reference picture in the paper-cut style (the picture above works), then paste a prompt. Size 16:9.
Scene 1: Shibuya crossing at dusk, Tokyo Tower behind
Paper-cut collage illustration in the exact style of the reference images: flat layered cut-paper shapes with soft drop shadows, torn and scissor edges, matte paper texture, muted documentary palette. Wide elevated view of Shibuya crossing in Tokyo at dusk: the scramble crossing with white zebra stripes in every direction, crowds of small paper people crossing, taxis and a green bus at the edges, mid-rise buildings with big screens and vertical signs, Tokyo Tower in red and white lattice rising behind, far skyline blocks, a large low sun disc and cut-paper cloud strips. Front left a girl with a red umbrella, back to camera. Light from the left, warm dusk. Palette: dusk lilac, peach, cream, tower red, neon teal. 16:9.
Scene 2: Senso-ji, the Kaminarimon gate and the pagoda at night
Paper-cut collage illustration in the exact style of the reference images: flat layered cut-paper shapes with soft drop shadows, torn and scissor edges, matte paper texture, muted documentary palette. Wide view of Senso-ji temple in Tokyo at night: the red Kaminarimon gate with its giant red paper lantern, the five-storey pagoda rising behind, temple roofs, Nakamise market stalls left and right with awnings, strings of small paper lanterns, small figures in kimono and yukata, a rickshaw, a cat on the stone path, a full moon disc. Right foreground a woman in a red kimono holds a paper parasol. Light from the lanterns, warm on cool night. Palette: midnight blue, lantern red, gold, cream, pine green. 16:9.
Scene 3: Mount Fuji over Lake Kawaguchi at dawn, torii and Shinkansen
Paper-cut collage illustration in the exact style of the reference images: flat layered cut-paper shapes with soft drop shadows, torn and scissor edges, matte paper texture, muted documentary palette. Wide view of Mount Fuji over Lake Kawaguchi at dawn: Fuji with its snow cap in two paper planes, forested hills, a white Shinkansen train crossing a bridge in the middle distance, the lake as flat blue paper planes with reflections, cherry trees in blossom on the banks, a red torii gate standing in the water, a big pale sun disc and cut-paper cloud strips. Front centre a fisherman in a small wooden boat. Light from the right, soft dawn. Palette: powder blue, blush pink, snow white, cream, torii red. 16:9.
Scene 4: Omoide Yokocho, a rainy neon alley in Shinjuku at night
Paper-cut collage illustration in the exact style of the reference images: flat layered cut-paper shapes with soft drop shadows, torn and scissor edges, matte paper texture, muted documentary palette. Wide view of Omoide Yokocho, a narrow rainy alley in Shinjuku, Tokyo, at night: layered alley facades left and right with vertical sign boards, neon signs in pink and cyan, rows of red paper lanterns, a ramen stall on the left with a counter, bowls and rising steam, two seated customers, small figures with umbrellas, puddles on the alley floor with neon reflections, thin rain streaks as paper strips, the tall Shinjuku towers behind. Left foreground the ramen cook ladles broth behind the counter. Light from the lanterns and neon, warm on cool night. Palette: charcoal, neon pink, cyan, lantern red, cream. 16:9.
2. The animation
Google Omni Flash 1.1 (Higgsfield, Google Flow or the API). Image to video, 16:9, 5 seconds, 720p. Start frame = a plain sky frame in the still's sky colours. End frame = the still. Rewrite the TIMELINE nouns with the layers in your still, back to front.
STYLE Paper-cut collage motion graphic. Every element in the attached image is a separate flat cut-paper layer with torn edges, a soft drop shadow and matte paper texture. Documentary title-sequence feel. Colours, shapes and composition stay exactly as in the attached image. NOT smooth 3D, NOT glossy, NOT photoreal. REFERENCE The attached image is the final frame. The video ends on this exact image, pixel-true, and holds. CAMERA: ONE STATIC SHOT, NO CUTS Locked-off camera, no zoom, no pan, no parallax. 16:9. TIMELINE (5 seconds) 00:00-00:00.3 Only the plain sky, nothing else on screen. 00:00.3-00:00.9 The sun or moon disc rises into place; the cloud strips slide in one at a time from the edges and settle. 00:00.9-00:01.6 The far layers rise from the bottom edge, farthest first, nearest last, each settling with a tiny paper wobble. 00:01.6-00:02.6 The buildings assemble piece by piece from back to front, each piece sliding in from below or the side and locking. 00:02.6-00:03.3 The mid-ground props slide in; the lanterns or lights switch on. 00:03.3-00:03.8 The small figures pop in one after another; the foreground hero element slides in last and locks. 00:03.8-00:04.9 LIVING HOLD. The composition is complete and nothing changes position. Only gentle secondary motion inside the finished picture (people take small steps, lanterns sway, water shimmers, rain falls). Paper-cut motion, small and calm. 00:04.9-00:05.0 Settle: identical to the attached image. AVOID Camera movement, cross-fades, morphing, elements changing shape or colour, extra elements not in the image, text changing, watermarks, music.
3. Put it together
- Cream background (#eeecdd), two 16:9 panels: the animation on top, the still below. Labels "Google Omni Flash" and "Chat GPT" in Lato Bold.
- Four scenes of 5 seconds, hard cuts, about 1.3 seconds of hold at the end.
- One music bed under all of it. We build it in code; any editor with two video tracks does the same.