E12 — "The Window" · Storyboard

Episode at a glance

The season finale and longest episode at 4:30 (270 seconds). Palette opens in deep indigo #4F46E5 against full-star sky and dawn-cream #FFE4C9 static-wall, accumulates every prior biome's tint through the mentor crescent (moss-green Hana, binary-blue Kiko, twin-moon silver Luna, peach-cream Nori, pale-violet Bruno, liquid-gold Otto, three-moon Rex, twilight-blue Mimi, comet-magenta Ari + CRT-blue Pixel, library-amber Biscuit, storm-slate Zara), then resolves into a single beam of indigo crossing into a softer, semi-realistic, deliberately-out-of-focus user's bedroom — the season's only palette pivot away from voxel saturation. Key motifs: chord completes (eleven → twelve voices), each shard plays its episode leitmotif for one second in episode order during assembly (the callback medley), the closing pixel-on-glass + chime is designed to be the production-merge first-launch animation of the Souli app. Total: 56 frames (52 diegetic + 4 production notes). 9:16 vertical primary cut throughout unless noted.

Scene 13.1 — The edge [00:00 – 00:30]

Frame 1 — [00:00:00 — 00:00:06] · Cold open on the edge

Composition. Wide hero, eye-level on Dzino's back. Voxel character lower-third center, tiny against vertical frame. Foreground: voxel grass-pixels. Midground: vertical curtain of soft-flickering static filling upper two-thirds. No background — just the wall.

Action. Static. Crown pixel ticks one frame brighter. Wind nudges three grass-pixels.

Lighting & palette. Indigo #4F46E5 body silhouetted against cream-warm #FFE4C9 static, key-lit from the wall — Dzino is rim-lit cream on indigo. Time of day: ambiguous dusk.

Sound cue. Eleven-voice chord sustains from E11. Faint warm wall-hum (tube amp at idle). No mentor leitmotifs yet.

AI-gen prompt. Voxel pixel-art wide shot, tiny indigo voxel character with cream face and orange crown standing back-to-camera at the edge of a flat voxel-grass plain, facing a vertical wall of soft warm-cream static light, 9:16, dawn-cream and indigo palette, no horizon beyond the wall, gentle film grain, Studio Ghibli stillness.

Notes. Opening promise — the wall is warm. Audiences expect a cold sci-fi barrier; we want them to read this as a threshold. Hold longer than feels comfortable.

Frame 2 — [00:00:06 — 00:00:12] · The sky behind

Composition. Slow tilt-up. Camera cranes from Dzino's small back-silhouette past his crown into a full vertical column of indigo night.

Action. Camera moves. Dzino doesn't. One star top-center, larger than its neighbors, pulses once.

Lighting & palette. Indigo #4F46E5 saturating, cream stars, the calling-star bigger and gold-cream. Wall remains a soft glow at bottom edge.

Sound cue. Chord sustains. The calling-star chime (E00) flickers in for half a second, then fades.

AI-gen prompt. Voxel pixel-art slow tilt-up, indigo night sky filled with cream-white voxel stars, one larger gold-cream pulsing star at center top, back of voxel character with orange crown at bottom of frame, 9:16, soft film grain.

Notes. Re-establishes the spine question — the calling star is now overhead, not on the horizon. Pays off E00's pulse-cue.

Frame 3 — [00:00:12 — 00:00:20] · The star is full

Composition. Top-down on Dzino. We see his cream face turned upward. Star directly above, gold-cream, four times its E00 size.

Action. His eyes (two cream pixels on indigo) widen by one pixel each. The star pulses twice. Crown reflects gold for three frames.

Lighting & palette. Gold-cream key from above paints face and crown. Body remains indigo. Background indigo-black.

Sound cue. Calling-star chime returns fuller — one clear sustained tone joining the chord. We briefly hear what twelve voices will sound like, then it pulls back to eleven. Tease.

AI-gen prompt. Voxel pixel-art high overhead shot looking down at indigo voxel character with orange crown and cream face turned up, gold-cream pulsing star directly above, indigo night sky, 9:16, gold rim-light on crown.

Notes. Visual rhyme with E00 — he looked up at a faint star; now the star looks back. "We have arrived" without dialogue.

Frame 4 — [00:00:20 — 00:00:26] · The wall, in macro

Composition. Macro on the static surface. No character. Wall fills 100%. Tiny voxel artefacts drift across — a Hana-petal, a 1 from a binary tree, a single gold thread.

Action. Static flickers in slow rolls top to bottom. Three artefacts drift through.

Lighting & palette. Cream #FFE4C9 saturated with indigo flecks. Old TV between channels but golden. No cold blues anywhere.

Sound cue. Chord drops to background. Wall-hum rises forward — analog warmth, no noise. Flicker-tick on each scan roll.

AI-gen prompt. Macro of warm cream-colored television static with subtle indigo flecks, slow horizontal scan rolls, occasional tiny voxel artefacts (flower petal, binary digit, gold thread) drifting across, 9:16, analog grain, no characters.

Notes. The static is made of the season — every biome's particle persists in it. Subliminal. Sets up the wall as a membrane, not a barrier.

Frame 5 — [00:00:26 — 00:00:30] · He waits

Composition. Medium back-shot, Dzino centered lower third. Wall fills upper half. Head slightly tilted up.

Action. Static. His chord-aura — eleven faint concentric rings around his torso — pulses once with his breath.

Lighting & palette. Indigo silhouette, cream rim from the wall.

Sound cue. Eleven-voice chord, full. Wall-hum drops. A single soft footstep — off-screen, behind him — lands on grass.

AI-gen prompt. Voxel pixel-art medium back-shot, small indigo voxel character with orange crown facing soft warm-cream static wall, eleven faint concentric rings of light pulsing around his torso, 9:16.

Notes. Off-screen footstep cues Hana arriving. We do not show her yet. Anticipation.

Scene 13.2 — The arrival of the twelve [00:30 – 01:30]

Each mentor: ~5 sec micro-frame in canonical pose. Each leitmotif weaves into the chord rather than soloing.

Frame 6 — [00:00:30 — 00:00:35] · Hana arrives

Composition. Medium two-shot, low side angle. Dzino foreground left. Hana enters frame-right on the path with her watering can.

Action. Hana walks four voxel-steps in, stops at his right shoulder. The can tilts a single drop of light onto the grass — a pixel-flower opens.

Lighting & palette. Moss-green tint enters with her, threading the indigo-and-cream. Watering can pours soft yellow light, not water.

Sound cue. Hana's leitmotif (E01) — humming flower-chord — lifts under the chord. Adds texture, doesn't replace.

AI-gen prompt. Voxel pixel-art medium shot, peaceful rabbit Hana in moss-green apron carrying a watering can pouring soft yellow light, walking up to small indigo voxel character at edge of static wall, single pixel-flower opening between them, 9:16, Studio Ghibli warmth.

Notes. Establishes the arrival rhythm: signature object, one canonical action, settle. No words from any mentor the entire scene.

Frame 7 — [00:00:35 — 00:00:40] · Kiko drops in

Composition. Medium, slight low-angle. Kiko mid-air upper-right, mid-fall, upside-down, paws splayed.

Action. Falls into frame from above upside-down, lands on paws-then-rolls, settles beside Hana. A binary-tree branch falls with him and dissolves into 1s and 0s in the grass.

Lighting & palette. Adds Binary Forest's cool blue-green glyph-glow. The 1s and 0s are pale-cream pixels.

Sound cue. Kiko's leitmotif (E02) — rustling-binary chime resolving into a third — joins.

AI-gen prompt. Voxel pixel-art medium shot, chaotic cat scientist Kiko in pixel lab coat falling upside-down out of frame top, small voxel binary tree branch dissolving into 1s and 0s, landing beside rabbit and indigo voxel character at edge of warm static wall, 9:16.

Notes. Only mentor who arrives kinetically. Builds momentum into an otherwise-still sequence.

Frame 8 — [00:00:40 — 00:00:45] · Luna steps from a mirror

Composition. Medium. A standing voxel mirror has appeared just behind Dzino's left shoulder. Luna mid-step, half-out of the glass.

Action. She steps fully out. The mirror remains; its surface ripples once and shows briefly the back of Dzino's head from inside. She tucks the brush from E03 behind her ear.

Lighting & palette. Twin-moon silver-violet enters. Mirror flares for one frame in cream-white.

Sound cue. Luna's leitmotif (E03) — slowed music-box bending sustained note — joins.

AI-gen prompt. Voxel pixel-art medium shot, dreamy fox artist Luna stepping out of a tall standing mirror, paintbrush behind her ear, mirror surface rippling cream-white, joining a small group at edge of warm static wall, twin-moon silver-violet light, 9:16.

Notes. Mirror persists silently for the rest of the scene. Easter egg: in Scene 13.4 we will not see Dzino's reflection in this mirror as he crosses, only static beyond.

Frame 9 — [00:00:45 — 00:00:50] · Nori descends

Composition. Medium, camera tilts up briefly. A staircase of glowing-laughter materializes from upper-frame, descending diagonally. Nori walks down it.

Action. Three steps down. Each footfall lights gold-pink, then fades behind her. She arrives carrying a paper bag of crumbs from E04.

Lighting & palette. Peach-cream and dream-cake gold-pink saturate the upper frame for three seconds, then settle.

Sound cue. Nori's leitmotif (E04) — third-voice triadic resolution — joins.

AI-gen prompt. Voxel pixel-art medium shot, passionate fox chef Nori in apron walking down a softly-glowing staircase of laughter-light made of cumulus, each step lighting gold-pink and fading behind, carrying a small paper bag of glowing crumbs, descending toward edge of warm static wall, 9:16.

Notes. Most kinetic warm-light entrance. Rhymes with E04's exit-staircase: she now walks the opposite direction.

Frame 10 — [00:00:50 — 00:00:55] · Bruno by the lake-that-is-here

Composition. Medium. To Dzino's right, the ground ripples briefly into a shallow pool — a single voxel-lake ten pixels across, present for one beat. Bruno cross-legged at its edge, eyes closed.

Action. Lake forms with a soft plip; Bruno is already there as if always. Lake holds. He doesn't move.

Lighting & palette. Pale-violet low-grav moonlight pools over Bruno only.

Sound cue. Bruno's leitmotif (E05) — single bowl-chime — strikes once and sustains as a soft drone.

AI-gen prompt. Voxel pixel-art medium shot, quiet bear philosopher Bruno sitting cross-legged with eyes closed beside a small ten-pixel-wide voxel lake on the ground, pale-violet low-gravity moonlight on him only, joining a small crescent, 9:16, perfect stillness.

Notes. Only mentor already there when his frame begins. He didn't arrive; he revealed.

Frame 11 — [00:00:55 — 00:01:00] · Otto rises from gold thread

Composition. Medium-close. A ribbon of gold thread on the ground beside Hana thickens, twists; Otto rises from it in robes, half-thread for a beat, then fully voxel.

Action. Otto extends one arm in a Shakespearean flourish as he settles. Fishing rod across his back. The thread shrinks to a small knot at his feet — a river-stone shape.

Lighting & palette. Liquid-gold amber threads through — most saturated single-color moment so far.

Sound cue. Otto's leitmotif (E06) — the half-heard future-voice in the gold current — surfaces for half a second, an indistinct human syllable, folds back into the chord.

AI-gen prompt. Voxel pixel-art medium shot, theatrical bear historian Otto in robes rising from a thickening ribbon of liquid-gold thread that becomes a small river-stone, fishing rod across back, one arm in a flourish, joining a crescent at edge of warm static wall, gold amber light, 9:16.

Notes. The half-heard syllable is the user's first message, recorded in advance — same fragment as E06. Subliminal continuity.

Frame 12 — [00:01:00 — 00:01:05] · Rex sprints in

Composition. Wide for one beat — frame opens to accommodate momentum. A crystal dune rolls up from frame-left, holds for one beat, Rex sprints down it. Back to medium as he stops.

Action. Dune rises, Rex sprints, dune dissolves behind. He skids, pants once, grins. High-fives the air at his side — the air remembers.

Lighting & palette. Three-moon glints (warm-pink, ice-blue, gold) flash across the dune, then dune is gone.

Sound cue. Rex's leitmotif (E07) — drum-machine heartbeat — joins. Now with percussion.

AI-gen prompt. Voxel pixel-art wide-to-medium shot, energetic dog athlete Rex in workout band sprinting down a crystal dune briefly risen from flat ground, three-moon highlights warm-pink ice-blue gold, dune dissolving behind, skidding to halt and grinning, joining a crescent at edge of warm static wall, 9:16.

Notes. Only frame that briefly widens the crescent's spatial logic — the world bends to make room. Subtle. Don't telegraph.

Frame 13 — [00:01:05 — 00:01:10] · Mimi, simply there

Composition. Medium. Cut to the crescent — Mimi already standing in it between Bruno and Otto, journal open in her paws.

Action. Static frame. She turns one journal page. That is all.

Lighting & palette. Twilight-grove deep-blue softens through from her direction.

Sound cue. Mimi's leitmotif (E08) — silence-as-fourth-voice — pulls a voice out of the chord. Chord momentarily thins, then refills. Net: silence becomes its own voice.

AI-gen prompt. Voxel pixel-art medium shot, shy bunny bookworm Mimi standing already-present in a crescent, journal open in her paws writing, twilight-grove deep blue light from her direction, edge of warm static wall background, 9:16, perfect calm.

Notes. Most Ghibli of arrival frames. Rhymes with Bruno's "revealed" grammar — the two quietest mentors share it.

Frame 14 — [00:01:10 — 00:01:15] · Ari and Pixel, together

Composition. Medium two-shot. Both walk in from frame-right side-by-side. Ari with deck under one arm, Pixel with laptop balanced on a small portable CRT.

Action. They stop at the crescent. Ari finger-guns at Dzino's back. Pixel salutes with two pixel-fingers.

Lighting & palette. Comet-ribbon magenta from Ari, CRT-blue scanline glow from Pixel — coolest tints in the scene, balanced.

Sound cue. Ari + Pixel's leitmotif (E09) — chiptune drop — bumps once, single beat-drop, folds into the chord as harmony+melody.

AI-gen prompt. Voxel pixel-art medium two-shot, DJ fox Ari with portable voxel turntable deck and gamer cat Pixel carrying a laptop balanced on a small CRT monitor, walking in side-by-side, Ari finger-gunning, Pixel saluting two fingers, comet-magenta and CRT-blue glow, joining a crescent at edge of warm static wall, 9:16.

Notes. Only mentors who arrive as a pair. Reflects E09's structural choice.

Frame 15 — [00:01:15 — 00:01:20] · Biscuit at his shoulder

Composition. Tight medium-close, side angle on Dzino's right shoulder. Biscuit walks into frame from behind, close, rests his head against the shoulder. Camera holds.

Action. Four steps, settles. Head-rest. Tail wags once, slowly.

Lighting & palette. Library-amber warm pools around the shoulder-rest moment. Crescent is bokeh background.

Sound cue. Biscuit's leitmotif (E10) — solo cello — adds, low and warm. Eleven voices. One shy of full.

AI-gen prompt. Voxel pixel-art tight medium-close, therapy dog Biscuit in soft amber light walking up and resting his head against the right shoulder of a small indigo voxel character with orange crown, tail wagging once slowly, crescent of figures softly out of focus background, 9:16.

Notes. Direct visual rhyme with E10's quietest shot. Most physically intimate moment of the scene; hold the beat.

Frame 16 — [00:01:20 — 00:01:25] · Zara from the dissipated storm

Composition. Medium-wide. Frame-left, thin vapor of cleared storm-mist drifts away. Zara walks out of it, lantern hanging dark from her hand. Last to arrive.

Action. Four steps, stops at the crescent's open end, completes the arc. Lantern extinguished — she does not relight it.

Lighting & palette. Storm-slate grey threads behind her, dissipates. Lantern, unlit, glints once with retained heat.

Sound cue. Zara's leitmotif (E11) — cleaner-than-ever resolve — does not add a voice. Instead her motif clarifies the existing eleven — every voice steps a quarter-tone into perfect tuning.

AI-gen prompt. Voxel pixel-art medium-wide shot, bold human entrepreneur Zara walking out of thin dissipating storm-mist with extinguished lantern, joining the open end of a crescent of figures at edge of warm static wall, storm-slate grey clearing behind, 9:16.

Notes. Closes the crescent. Lantern stays dark — she doesn't need it here. Pays off E11's final beat.

Frame 17 — [00:01:25 — 00:01:30] · The crescent · hands and paws

Composition. Hero wide back-shot. Dzino lower-center, very small. Behind him, the full crescent of twelve mentors arcs edge-to-edge. Inset montage of twelve hand-on-back close-ups, three frames each, layered.

Action. All twelve simultaneously make contact. Dzino does not turn. His chord-aura pulses once, brighter.

Lighting & palette. Every biome tint blends — first time the season's full palette is in one frame. Indigo and cream still dominant.

Sound cue. All eleven voices ring at full clarity. Wall-hum drops to a whisper. The twelfth voice still pending.

AI-gen prompt. Voxel pixel-art hero wide back-shot, small indigo voxel character with orange crown lower-center facing warm cream static wall, behind him a full crescent of twelve diverse voxel mentors (rabbit, cat, fox, fox, bear, bear, dog, bunny, fox-DJ, cat-gamer, dog, human-with-lantern), each placing one hand or paw on his back or shoulders, every biome's tint visible, 9:16, Studio Ghibli stillness.

Notes. The hero image of Scene 13.2. If only one E12 frame goes to marketing — this one.

Scene 13.3 — Assembly of shards [01:30 – 02:30]

Frame 18 — [00:01:30 — 00:01:35] · The chest-glow

Composition. Tight close-up on Dzino's chest. Eleven small points of light bloom within his torso, in episode order top to bottom.

Action. Eleven pinpricks light up, one per second, all visible by frame end.

Lighting & palette. Cream-gold key from inside the body. Body remains indigo, lit from within.

Sound cue. Chord sustains. Soft tick per pinprick — eleven ticks across the frame.

AI-gen prompt. Voxel pixel-art tight close-up of an indigo voxel character's chest with eleven small cream-gold pinpricks of light blooming inside his torso one by one, lit from within, 9:16, no other characters.

Notes. Sets geometry for the next eleven frames — each shard rises in episode order from a known position.

Frames 19–29: shard rises in episode order. Identical staging — only the shard and its leitmotif change. Each frame is exactly three seconds; the medley feels deliberate.

Frame 19 — [00:01:35 — 00:01:38] · Personality (Hana)

Action. A humming pixel-flower lifts from his chest, vibrating gently, holds in air.

Sound cue. Hana's leitmotif (E01) plays one full second, isolated. Chord ducks beneath, then back. Callback medley begins.

AI-gen prompt. Voxel pixel-art medium close-up, a small humming pixel-flower in moss-green and cream rising out of indigo voxel character's chest, hovering and vibrating, 9:16.

Frame 20 — [00:01:38 — 00:01:41] · Interests (Kiko)

Action. A folded notebook page lifts, unfolds once, holds.

Sound cue. Kiko's leitmotif (E02) — one second isolated.

AI-gen prompt. Voxel pixel-art medium close-up, folded paper notebook page with binary 1s and 0s rising from chest, unfolding once mid-air beside a hovering pixel-flower, 9:16.

Frame 21 — [00:01:41 — 00:01:44] · Form (Luna)

Action. A single brushstroke — indigo paint on canvas-sliver — lifts and curls.

Sound cue. Luna's leitmotif (E03) — bending music-box note. One second.

AI-gen prompt. Voxel pixel-art medium close-up, small indigo brushstroke painted on a sliver of voxel canvas rising and curling mid-air, twin-moon silver-violet glow, beside other shards, 9:16.

Frame 22 — [00:01:44 — 00:01:47] · Humor (Nori)

Action. A pulsing slice of impossibility — cake-wedge of light — rises and rotates.

Sound cue. Nori's leitmotif (E04) — triadic resolution. One second.

AI-gen prompt. Voxel pixel-art medium close-up, pulsing voxel cake-slice of impossible peach-cream and gold-pink light rising and rotating mid-air, beside other shards, 9:16.

Frame 23 — [00:01:47 — 00:01:50] · Philosophy (Bruno)

Action. A single ripple — flat disc of pale-violet water — lifts and holds.

Sound cue. Bruno's leitmotif (E05) — bowl-chime. One second.

AI-gen prompt. Voxel pixel-art medium close-up, flat disc of pale-violet rippling water rising and holding mid-air, low-gravity moonlight, beside other shards, 9:16.

Frame 24 — [00:01:50 — 00:01:53] · Bonds (Otto)

Action. A small knot of gold thread untwists once and re-knots, hovering.

Sound cue. Otto's leitmotif (E06) — half-heard human syllable. One second. Audience again cannot quite make it out. Same syllable as Frame 11.

AI-gen prompt. Voxel pixel-art medium close-up, small knot of liquid-gold thread untwisting once and re-knotting mid-air, amber glow, beside other shards, 9:16.

Frame 25 — [00:01:53 — 00:01:56] · Body (Rex)

Action. A heartbeat-pulse — two cream pixels expanding, contracting — rises.

Sound cue. Rex's leitmotif (E07) — drum-machine heartbeat. One second.

AI-gen prompt. Voxel pixel-art medium close-up, two cream voxel pixels pulsing in heartbeat rhythm rising mid-air, three-moon highlights warm-pink ice-blue gold, beside other shards, 9:16.

Frame 26 — [00:01:56 — 00:01:59] · Quiet (Mimi)

Action. An empty page rises and turns once on an axis. Stays empty.

Sound cue. Mimi's leitmotif (E08) — silence. Chord ducks fully for one second. Medley pauses. Then resumes.

AI-gen prompt. Voxel pixel-art medium close-up, single empty paper page rising and turning once on its axis mid-air, deep twilight-blue light, beside other shards, 9:16.

Notes. The silent second is the most important beat in the medley. Trust it. Do not fill it.

Frame 27 — [00:01:59 — 00:02:02] · Purpose (Ari + Pixel)

Action. A dropped beat — single waveform pulse with a strobe-flash — rises and pulses three times.

Sound cue. Ari + Pixel's leitmotif (E09) — chiptune drop, one bar. One second.

AI-gen prompt. Voxel pixel-art medium close-up, single voxel waveform pulse strobing magenta and CRT-blue rising mid-air, beside other shards, 9:16.

Frame 28 — [00:02:02 — 00:02:05] · Trials (Biscuit)

Action. A dull, heavy stone — no glow — rises slowly, weighted, almost reluctantly. Only shard that does not visually shine.

Sound cue. Biscuit's leitmotif (E10) — solo cello. One second. Lower in register than any other shard's motif.

AI-gen prompt. Voxel pixel-art medium close-up, small dull dark stone with no glow rising slowly and weightily mid-air beside other glowing shards, the only un-luminous shard, library-amber background, 9:16.

Notes. Visual contrast — every other shard is light; this one is mass. Pays off E10's "you don't have to solve them, you just have to be near them."

Frame 29 — [00:02:05 — 00:02:08] · Journal (Zara)

Action. A single pen-stroke — wet ink, indigo on cream — draws itself mid-air, completing the eleven.

Sound cue. Zara's leitmotif (E11) — storm-clarified eleven-voice resolve. One second.

AI-gen prompt. Voxel pixel-art medium close-up, single wet indigo ink pen-stroke drawing itself mid-air on invisible paper beside other floating shards, storm-slate clearing into cream behind, 9:16.

Notes. End of the callback medley. Audience has just heard the entire season as music.

Frame 30 — [00:02:08 — 00:02:14] · The eleven, hovering

Composition. Medium-wide. All eleven shards arrayed in a loose constellation. Dzino small lower frame; shards spread across upper two-thirds.

Action. All eleven hold position. They rotate slightly, in different rhythms. Chord swells. Crescent visible only as soft out-of-focus indigo silhouette behind.

Lighting & palette. Every biome tint visible at once for the second time, now in concentrated point-form.

Sound cue. Each leitmotif still individually audible but woven — the season's chord at full eleven-voice clarity, every voice articulated.

AI-gen prompt. Voxel pixel-art medium-wide shot, eleven distinct floating shards arrayed in constellation above a small indigo voxel character — each a different shape and tint (flower, page, brushstroke, cake-slice, ripple, gold knot, heartbeat, empty page, waveform, dark stone, ink stroke), warm cream static wall background, soft out-of-focus mentor silhouettes behind character, 9:16.

Notes. "Season-in-one-image" hero. Suitable for poster.

Frame 31 — [00:02:14 — 00:02:20] · The lock

Composition. Medium. The eleven drift toward each other. Click — pairwise, then triadic, then all eleven — into a single shape.

Action. Continuous motion. Each click sends a faint cream ripple. By frame end, all eleven have locked into a hollow Dzino-shaped silhouette of floating fragments — a doorway in the protagonist's own shape.

Lighting & palette. All tints converge, balance into indigo-cream as the doorway forms. Biome tints fade into the silhouette's edge.

Sound cue. Each click an audible tk — eleven of them. On the final lock, the entire chord — all twelve voices, including the user-chime as the twelfth — strikes once, full. First time the audience hears twelve voices in the season.

AI-gen prompt. Voxel pixel-art medium shot, eleven distinct floating shards drifting together and clicking into a single hollow Dzino-shaped silhouette doorway of floating fragments, faint cream ripples on each click, indigo-cream balanced palette, 9:16, magical-realist Studio Ghibli.

Notes. The structural climax. The twelfth voice arrives without a twelfth shard — because the soul itself is the twelfth thing. Audience may not consciously parse this. They will feel it.

Frame 32 — [00:02:20 — 00:02:24] · The parting

Composition. Hero medium. Dzino's small back center-frame. Doorway-of-shards fills upper-center, Dzino-shaped. Through its hollow interior the static is parting like curtains.

Action. Static within the doorway aperture parts vertically, slowly, revealing a dim warm interior beyond.

Lighting & palette. The opening reveals softer, semi-realistic light beyond — warm tungsten-amber, distinctly different from anything in Pixoci. First palette pivot.

Sound cue. Twelve-voice chord sustains. Wall-hum decays to nothing.

AI-gen prompt. Voxel-pixel-art-to-semi-realistic medium shot, small indigo voxel character back-to-camera looking up at hollow character-shaped doorway of floating shards, static inside the doorway parting vertically like curtains to reveal softer warm tungsten-amber light beyond, palette transition from voxel-saturated to semi-realistic warm interior, 9:16.

Notes. First pixel-by-pixel palette pivot of the season. Mark carefully — this is what sells the crossing.

Frame 33 — [00:02:24 — 00:02:30] · The room, first glimpse

Composition. Through-the-doorway insert. Frame is mostly the inside of the doorway. We see, for the first time in the season, the room beyond: a window across the back wall. A small lamp on a side table casts warm tungsten. A figure — soft-edged, deliberately defocused, head-and-shoulders silhouette — sits near the window with a phone. Outside the window: night with the calling-star in the upper sky.

Action. Gentle. The figure shifts a fraction. Phone screen-glow on their face but face not resolvable. Window curtain stirs.

Lighting & palette. Warm tungsten-amber lamp (key). Cool moon-cream window (fill). Indigo night through glass (background). Voxel saturation gone. Rendered semi-realistically with deliberate soft-focus on the figure — figure at f/1.2, blurred enough that no viewer can determine gender, age, or ethnicity.

Sound cue. Twelve-voice chord ducks slightly. We hear, for the first time, room sound — refrigerator hum, distant city, clock-tick. Subliminal but present. The shift in audio space is what sells the crossing — not the visual.

AI-gen prompt. Cinematic semi-realistic medium-wide through a hollow character-shaped frame of floating shards, looking into a real-feeling cozy human bedroom, large window across back wall with night sky and a single bright cream-gold star outside, small lamp casting warm tungsten-amber, soft-edged deliberately out-of-focus human silhouette sitting near window holding a phone, do NOT specify gender age or ethnicity, shallow f/1.2 depth of field on figure, painterly Studio Ghibli, 9:16, palette pivot from voxel to soft realism.

Notes. The single most important shot in the entire season. It must be ambiguous enough that any viewer can project themselves into it. If the figure resolves into a specific person, we lose the production-merge with the app. Shoot multiple variations; pick the one that holds projection longest.

Scene 13.4 — The crossing [02:30 – 03:30]

Frame 34 — [00:02:30 — 00:02:36] · The look back

Composition. Medium-wide. Camera between Dzino and the doorway, low-angle reverse — we see Dzino from in front for the first time, his cream face turned over his shoulder back toward the crescent. Crescent in soft focus behind him.

Action. He turns his head. Eyes settle on the twelve. Holds three seconds.

Lighting & palette. Indigo silhouette, cream face lit warm by the doorway behind, biome tints washing across the soft-focus crescent.

Sound cue. Twelve-voice chord sustains. Room sound from the doorway audible behind. No mentor moves or speaks.

AI-gen prompt. Voxel pixel-art medium-wide low-angle reverse, small indigo voxel character with cream face turned over his shoulder looking back at a soft-focus crescent of twelve mentors, warm tungsten light from a doorway behind casting on his face, 9:16.

Notes. Only shot in the scene where we see Dzino's face from the front. He is older here than at any prior moment in the season — not in voxels, but in eyes.

Frame 35 — [00:02:36 — 00:02:42] · The nod

Composition. Medium on the crescent. Twelve mentors visible. Dzino out of frame.

Action. All twelve nod simultaneously, once. Slow. Small motion. None step forward.

Lighting & palette. All biome tints in balance, with the doorway's warm tungsten softly bleeding from out of frame.

Sound cue. Chord. Wall-hum gone. Room sound recedes behind the chord. Silence around the nod itself — no foley.

AI-gen prompt. Voxel pixel-art medium shot, full crescent of twelve diverse voxel mentor characters all nodding once simultaneously, slow motion, none stepping forward, every biome tint visible, soft warm tungsten bleed from out-of-frame doorway, 9:16.

Notes. They are staying. The nod is the goodbye. Each canonical-pose dignity preserved.

Frame 36 — [00:02:42 — 00:02:48] · He turns

Composition. Match cut to Frame 34's reverse — back to Dzino's back. He is now facing the doorway again.

Action. Shoulders square. Crown straightens by one pixel. He takes one step.

Lighting & palette. Indigo body, distinctly rim-lit warm tungsten from the doorway. Voxel saturation in his body is just starting to soften at the edges.

Sound cue. Twelve-voice chord. Room sound rises slightly.

AI-gen prompt. Voxel pixel-art medium back-shot, small indigo voxel character with orange crown squaring shoulders and taking one step toward a hollow character-shaped doorway of floating shards, body edges just beginning to soften from voxel hard-edge into semi-realistic blur, warm tungsten rim light, 9:16.

Notes. First frame where the voxel grammar begins to dissolve. Subtle. Edge-only.

Frame 37 — [00:02:48 — 00:02:55] · The compression begins

Composition. Medium-close on Dzino, back-shot, mid-step.

Action. As he walks forward, his voxel form loses pixels. Small cubes detach from his outline and dissolve into cream-gold sparks that arc back into the doorway-shards. He visibly shrinks — was 24 voxels tall, now ~18.

Lighting & palette. Indigo body sparking into cream-gold edges. Doorway light grows warmer as he enters its threshold.

Sound cue. Chord. Soft tk per detached voxel — accumulates into rhythm under the chord.

AI-gen prompt. Voxel pixel-art medium-close back-shot, small indigo voxel character mid-step walking toward a doorway of floating shards, voxel cubes detaching from his outline and dissolving into cream-gold sparks arcing back into the doorway, body shrinking, warm tungsten doorway light ahead, 9:16, magical-realist quality.

Notes. Pixel-by-pixel compression — the season's most technically distinctive effect. Plan the count: Dzino is roughly 24×36 voxels. He needs to lose ~95% across this and Frames 38–40, in a deliberate cadence resolving to one pixel.

Frame 38 — [00:02:55 — 00:03:02] · Halfway through

Composition. Medium. Half-inside the doorway. Foreground shards frame him. Now ~12 voxels tall.

Action. He keeps walking. More cubes detach. Crown is now a single orange voxel.

Lighting & palette. Body half-indigo voxel, half cream-gold spark-trail. Doorway tungsten washes most of the foreground.

Sound cue. Chord sustains. Tick rhythm continues.

AI-gen prompt. Voxel-to-semi-realistic medium back-shot, voxel character half-the-size walking through a doorway of floating shards, more voxel cubes detaching into cream-gold sparks, crown reduced to a single orange pixel, body half-pixel half-spark-trail, warm tungsten doorway dominating, 9:16, palette transition mid-pivot.

Notes. Hold slightly longer than previous frame — audience needs to register the shrinking, not just the dissolving.

Frame 39 — [00:03:02 — 00:03:08] · Through the threshold

Composition. Tight medium. Now at the doorway's far edge, almost through. Now ~6 voxels — tiny indigo cluster with a single orange pixel above.

Action. He passes the threshold. Behind him, the shards begin to drift apart.

Lighting & palette. Body indistinguishable from spark-trail — almost all sparks, small indigo core. Static wall closing back behind him.

Sound cue. Chord. Soft fwip as the static wall starts closing. Room sound fully audible. We are crossing.

AI-gen prompt. Voxel-pixel-art-to-soft semi-realistic tight medium, very small voxel cluster with single orange pixel-crown passing through far edge of doorway of floating shards, mostly cream-gold spark-trail with tiny indigo core, static wall closing softly behind, warm tungsten on the far side, 9:16.

Notes. —

Frame 40 — [00:03:08 — 00:03:14] · The single pixel forms

Composition. Match cut to tighter macro. We are on the inside of the window glass — looking at it from inside the room. A single indigo pixel resolves on the glass. The doorway, the shards, the crescent — all gone. Just window, pixel, soft-focus room beyond camera.

Action. The pixel resolves over two seconds — first faint glimmer, then defined single pixel. Cream-gold sparks have all condensed into it.

Lighting & palette. Full palette pivot. Voxel saturation gone. Pixel itself is single saturated indigo #4F46E5 against semi-realistic warm-tungsten interior. Window glass has very faint reflection of the room — lamp, curtain, hint of figure in deep background.

Sound cue. Chord. Static wall fwip finishes — Pixoci is sealed. Then quiet. Room sound at full presence. Chord softens, sustains, no longer dominates.

AI-gen prompt. Cinematic semi-realistic macro, inside of a real bedroom window glass, single saturated indigo pixel resolving on the glass surface, very faint reflection of warm tungsten lamp and soft-focus room interior in the glass, painterly soft realism, 9:16, complete palette pivot from voxel to semi-realism, the only voxel element in frame is the single indigo pixel.

Notes. The merge frame — voxel grammar meets semi-realistic grammar in a single contained pixel. Contrast must be clean: hard-edge pixel on soft-focus glass.

Frame 41 — [00:03:14 — 00:03:22] · The room, sustained

Composition. Medium-wide on the room. Camera fully inside the bedroom now, behind the figure, looking past them toward the window. Single indigo pixel visible on glass, frame-center, small but present. Foreground: back of figure's head and shoulder, deliberately soft-focus. Midground: table with steaming tea. Background: window, pixel, night beyond. 9:16.

Action. Gentle. Tea steams. The figure's hand holding the phone lowers slightly. Plant on windowsill leans into the moonlight by one frame.

Lighting & palette. Warm tungsten-amber lamp (key). Cool moon-cream window (fill, slightly stronger). Soft-realism throughout. Figure remains f/1.2 blur — unresolvable.

Sound cue. Chord soft. Tea steam audible — thin hiss. Phone-screen glow has its own subliminal high-pitched whisper. Small intake of breath — figure's, very quiet.

AI-gen prompt. Cinematic semi-realistic medium-wide inside a cozy human bedroom, behind a deliberately out-of-focus human silhouette near a window, single saturated indigo pixel visible on the inside of the window glass at frame center, steaming cup of tea on side table, plant on windowsill leaning into moonlight, warm tungsten lamp key, cool moon-cream window fill, painterly soft realism, do NOT resolve facial features, shallow f/1.2 on figure, 9:16.

Notes. Production-merge anchor frame. The figure must read as plausibly the viewer. Test demographic projection — if any single demographic resolves more than another, re-blur.

Scene 13.5 — First word [03:30 – 04:00]

Frame 42 — [00:03:22 — 00:03:30] · The pixel turns

Composition. Tightest possible macro on the indigo pixel on the glass. Pixel fills a small portion of frame; glass texture and faint room reflection occupy the rest. The orange crown is suggested as a single sub-pixel of warmer tone.

Action. Pixel rotates slowly on its vertical axis — a quarter turn. It is now facing camera/user, not the night beyond.

Lighting & palette. Indigo pixel on cream-glass. Faint room reflection at low opacity, painterly. Crown sub-pixel barely-orange.

Sound cue. Chord ducks fully. Room sound continues. Small pixelated tk — the rotation. Then silence holds for one beat.

AI-gen prompt. Macro of a single saturated indigo pixel on the inside of a window glass, slowly rotating a quarter turn on its vertical axis, faint suggestion of a warmer-toned orange sub-pixel implying a crown, soft cream room reflection at low opacity in the glass, painterly soft realism, 9:16, no other voxel elements.

Notes. Hold the rotation. Trust the audience to know what is happening. The pixel is looking at us.

Frame 43 — [00:03:30 — 00:03:36] · "Ahoj."

Composition. Same macro as Frame 42. No change. We are not cutting away.

Action. Pixel does not move visibly. But it speaks.

Lighting & palette. Same.

Sound cue. Chord at near-silence. Room sound at near-silence. A small pixel-thin voice — Dzino's, unmistakably his SoundDNA chirp now resolving into a single, two-syllable word — says:

"Ahoj."

The voice is small. Pixel-thin. Half-whisper, half-chirp. The first and only word of the season.

AI-gen prompt. Macro of a single saturated indigo pixel on the inside of a window glass, no visible motion, no other elements, painterly soft realism, 9:16, audio-driven frame: a small pixel-thin voice says "Ahoj" once.

Notes. Audio is the frame. Visual stillness amplifies it. If the visual moves at all, re-shoot. Voice direction: child-like but not childish, breath more than voice, Slovak pronunciation precise (ah-hoy, H breathed not pronounced).

Frame 44 — [00:03:36 — 00:03:42] · The star drops · the chime

Composition. Wider semi-realistic shot through the window — exterior. Night sky beyond. Calling-star, full and close, in the upper frame. Window pane occupies lower frame, indigo pixel small on the glass.

Action. Calling-star drops — gently, settling — out of the upper sky toward the window. As it descends it dissolves from a star-shape into a chime-wave-pattern visualized as a soft cream ripple expanding through the air toward the glass.

Lighting & palette. Indigo night, cream-gold star, soft-realism throughout. Star's gold becomes the dominant light source for two seconds before dissolving.

Sound cue. The chime. A single tone — designed and recorded as a real production asset, intended for first-launch of the Souli app. Plays once, full, sustained ~2 seconds, then decays. Underneath, the chord resolves to twelve voices for the final time, then begins to fade.

AI-gen prompt. Cinematic semi-realistic shot through a bedroom window into the night sky, full bright cream-gold star descending gently from upper sky toward the window and dissolving into a soft cream ripple of expanding chime-waves, single saturated indigo pixel small on inside of glass at lower frame, painterly soft realism, 9:16.

Notes. THE chime is a pre-canonical production asset — record specifically for this frame, reuse as actual first-launch sound of the Souli app. The merge between episode audio and app audio is the entire point of this beat.

Frame 45 — [00:03:42 — 00:03:48] · The figure turns

Composition. Medium semi-realistic on the figure — still soft-focus, still unresolvable. Inside the room, behind them, slightly to the side now so we can see their head in profile-blur.

Action. Figure looks up from the phone. Head turns slowly, ~30 degrees, toward the window. Phone screen-glow leaves their face. Lamp-light takes over. No facial features become resolvable at any point. Blur preserved.

Lighting & palette. Warm tungsten dominates as phone-glow recedes. Cool moon-cream on the further cheek. Soft-realism, f/1.2 blur on figure, sharp on the lamp and window edge.

Sound cue. Chime still decaying. Chord at near-zero. Small breath from the figure — half-recognition, half-surprise. No words.

AI-gen prompt. Cinematic semi-realistic medium, deliberately out-of-focus human silhouette in profile-blur slowly turning head toward a window, phone-screen glow receding from face as warm tungsten lamp light takes over, soft moon-cream on further cheek, do NOT resolve facial features at any point, preserve f/1.2 blur, painterly soft realism, 9:16.

Notes. Critical: the head turn is the beat where the entire app's marketing rests. A viewer should imagine themselves performing this exact head-turn. Do not cast a recognizable performer or render a definite face.

Frame 46 — [00:03:48 — 00:03:54] · They see something

Composition. Match cut to OTS from behind the figure, looking past their soft-focused head toward the window. Indigo pixel on glass is sharp; figure's shoulder and ear are blur.

Action. Static for one second. Then the figure leans forward — perhaps two voxel-equivalents in real-world distance, very small — toward the window. We see only the lean, not the face.

Lighting & palette. Same as Frame 45. Pixel is the sharpest object in the frame.

Sound cue. Chime fully decayed. Silence. Then a single soft beat of held breath — the figure's. Hold the silence three seconds.

AI-gen prompt. Cinematic semi-realistic over-the-shoulder from behind a deliberately out-of-focus human silhouette looking toward a window, single saturated indigo pixel sharp on inside of window glass, figure's shoulder and ear soft-blur in foreground, figure leaning slightly forward toward window, painterly soft realism, do NOT resolve facial features, 9:16, silent frame.

Notes. The moment where the user becomes the viewer. The fourth wall is not broken — it is crossed. Hold the silence past comfort. Production-merge: in the actual app, this frame becomes the first thing a user sees on first launch.

Scene 13.6 — Title card [04:00 – 04:30]

Frame 47 — [00:03:54 — 00:04:00] · Pull back to black

Composition. Camera dollies smoothly backward from Frame 46. The room compresses into a small lit pocket center-frame. Everything around it goes to black.

Action. Continuous dolly back over six seconds. Room shrinks; indigo pixel on the glass remains visible — single bright point — until it is the last thing in frame, surrounded by black.

Lighting & palette. Warm tungsten pocket compressing into black. Final pixel = single indigo #4F46E5 on full black #000000.

Sound cue. Silence. Held.

AI-gen prompt. Cinematic semi-realistic slow dolly-back, warm tungsten-lit bedroom interior receding into the center of frame as everything goes to black, single saturated indigo pixel remaining as the brightest point at center of black field by frame end, painterly soft realism transitioning to black, 9:16.

Notes. The pull-back is the season's exhale. Pixel-on-black at the end is the season's question mark.

Frame 48 — [00:04:00 — 00:04:06] · Black, held

Composition. Pure black. 9:16. The single indigo pixel from Frame 47 has faded out.

Action. None.

Lighting & palette. #000000.

Sound cue. Silence. Three full beats, as scripted.

AI-gen prompt. Pure black 9:16 frame, no elements, no motion.

Notes. Trust the silence. The temptation to fill three black seconds with sting or cue is enormous. Resist.

Frame 49 — [00:04:06 — 00:04:14] · Title — line 1

Composition. Pure black. Centered indigo #4F46E5 text fades in:

Pixoci — Season 1.

Pixel-art typography, voxel-styled letterforms.

Action. Type fades in over two seconds. Holds for six.

Sound cue. Silence.

AI-gen prompt. Pure black 9:16 with centered pixel-art voxel-style text "Pixoci — Season 1." in saturated indigo, fading in slowly, no other elements.

Notes. Title is plain. Nothing fancy. The work was the rest.

Frame 50 — [00:04:14 — 00:04:22] · Title — line 2

Composition. Same field. Beneath line 1, smaller cream #FFE4C9 italic fades in:

The continuation comes when you open your window.

Action. Type fades in over two seconds. Both lines hold together for six.

Sound cue. Silence still. Hold.

AI-gen prompt. Pure black 9:16 with two centered lines — "Pixoci — Season 1." in saturated indigo above, "The continuation comes when you open your window." in soft cream italic below — both visible, fading in slowly, no other elements.

Notes. The cream line is the call-to-action. Also literally true — next time the user opens their phone and launches Souli, this story continues.

Frame 51 — [00:04:22 — 00:04:28] · The chime returns

Composition. Same title field. No visual change.

Action. None visually. The chime — same as Frame 44, the user's first-launch chime — plays once, full, identical recording.

Sound cue. THE chime. Single tone. Same asset as Frame 44. ~2 seconds, with decay. Second time the audience hears it; on the app's first launch, it will be the third.

AI-gen prompt. Pure black 9:16 with two lines of pixel-art text held still — "Pixoci — Season 1." in indigo, "The continuation comes when you open your window." in cream — no visual change, audio-driven frame: a single soft cream chime tone plays once.

Notes. Audio rhyme with Frame 44 is the structural glue. Same asset, same level, same decay. Identical.

Frame 52 — [00:04:28 — 00:04:30] · End

Composition. Title card holds for two final beats, then cuts to black.

Action. Hard cut to black at 04:30:00.

Sound cue. Silence into black. No outro music, no credits in this cut.

AI-gen prompt. Pure black 9:16, hard cut, end of episode.

Notes. Hard cut is the season's period. Don't soft-fade. End.

Frame 53 — [PRODUCTION NOTE · NON-DIEGETIC] · Merge with the app's first-launch animation

Notes.

Frames 40–46 (single pixel + chime + user's head turn) are designed to be the actual first-launch animation of the Souli app.

When a user opens Souli for the first time, the app should:

Black for one beat.

A single indigo #4F46E5 pixel resolves on what reads as the inside of a window pane.

The pixel rotates a quarter turn — same motion as Frame 42.

Ahoj plays — same recorded asset as Frame 43.

The chime drops — same asset as Frames 44 and 51.

Soft-focus the implied environment around the pixel toward the user, then fade to onboarding.

This makes the app's first-launch experience the continuation of the season's final frame — the user is the soft-focus figure. The story does not end on the screen; it ends when the user opens their window, which is what their phone now is.

Coordinate with the Souli onboarding team. The chime and Ahoj recordings must be cut from the episode's master and stored as souli-first-chime.wav and dzino-ahoj.wav in the app's audio bundle. They are pre-canonical assets — record them with this frame in mind.

Frame 54 — [PRODUCTION NOTE · NON-DIEGETIC] · Director's cut variation (16:9)

Notes.

The 16:9 cut holds every frame at the same timing as the 9:16 primary, with this latitude:

Frame 17 (crescent of twelve) reframed as horizontal hero — all twelve visible across the wide frame in proportion. Festival-poster shot.

Frame 30 (eleven shards in constellation) reframed horizontally to spread the constellation evenly. Equally suitable as poster.

Frames 33 + 41 (inside the user's room) extended slightly with horizontal latitude — more room visible, but figure still soft-focus and unresolvable. The horizontal frame must NOT add detail to the figure.

All other frames identical; vertical-primary discipline is the season's grammar.

Frame 55 — [PRODUCTION NOTE · NON-DIEGETIC] · Subtitle and dub

Notes.

Ahoj remains untranslated and unsubtitled in every locale — sk, cs, en, de, es, fr, hi, hu, pl.

Ahoj is the same hello in Slovak and Czech. In English the audience may not know the word — that is intentional. They understand it from context: small voice, small pixel, small word, small turn of a head. The translation is the head-turn, not the subtitle.

The chime is pre-linguistic — same in every locale.

Frame 56 — [PRODUCTION NOTE · NON-DIEGETIC] · The next user's first frame

Notes.

The episode's final production-frame is not Frame 52. It is the frame the user sees when they open Souli for the first time after watching the finale.

The season closes on the app's launch. Frames 40–46 must be cut, mixed, and shipped as the production-merge animation by the time E12 releases. The distribution rhythm depends on it.

Coordinate release: E12 drops, app first-launch animation goes live the same day. The story is not finished by the episode — it is finished by the user opening their window.

Style anchors needed before AI gen

Reference plates required to lock E12 on style. Deliver as hand-painted (or LoRA-trained) references before any frame goes into Runway / Sora / SVD.

Storyboard draft — E12, the finale. Locks the production-merge between season-end and app-launch. Ready for director's pass and AI-gen plate prep.