AI Paper Animation · OpenArt Workflow and Complete Prompt Playbook

Paste this word for word under every image and video prompt.

Handmade paper cutout animation. Every element in the frame is cut from real physical paper and card stock and arranged in stacked flat layers on a shallow tabletop set. Visible paper grain and fibre texture across every surface. Torn and scissor-cut edges with slight fray and fibre lift, never clean vector edges. Each layer casts a small soft contact shadow onto the layer behind it, reading real separation in depth of a few millimetres. Matte uncoated stock throughout, absolutely no gloss, no plastic sheen, no specular highlight anywhere. Colour comes from the paper itself, dyed and printed stock, slightly uneven in saturation across a single sheet the way real card is. Small physical imperfections present and welcome: a fingerprint smudge, a faint fold crease, a fibre catching light at an edge, a hairline gap where two pieces meet. Lit by one soft broad source from upper camera-left, warm and even, with gentle falloff across the tabletop and no hard shadow. Photographed from directly overhead on a real camera with a normal prime, shallow but not extreme depth of field, fine natural film grain across the whole frame. Real paper photographed on a real table, never a 3D render of paper, never a digital illustration imitating paper, never a flat vector graphic.

Add this word for word to video prompts only, after the style lock.

All movement is physical paper movement. Pieces slide across the plane, pivot at visible paper joints, hinge and fold along real crease lines, and are lifted and swapped in and out of the frame. Layers stay flat and parallel to the table at all times and never rotate in three dimensions, never turn to reveal a side or a thickness, never bend as if made of rubber. Contact shadows travel with their piece, staying locked to the piece that casts them. Transitions between states happen by pieces physically sliding away, folding closed, or being covered by another piece, never by dissolve, never by morph, never by fade.

Use text-to-video at 1080p for 30 seconds, then append both universal blocks.

An explainer about how a zipper works, built from paper as one macro tabletop set: a long vertical zipper of two ochre kraft fabric tapes, two rows of small charcoal paper teeth, and one large chalk-cream slider piece, with cut paper labels that enter and exit. All lettering is cut from charcoal card with torn edges, never a clean digital typeface, spelled exactly and kept flat, whole, and legible: TEETH, then SLIDER, then LOCKED. Every piece is free-standing and attached to nothing beyond the zipper itself, no arm, no rod, no connector anywhere.

0-6s: the two kraft tapes slide in from the frame sides and settle parallel with a gap between them, each carrying its row of charcoal teeth angled outward; cut letters spelling TEETH slide in along the bottom.
6-12s: the cream slider piece slides in from the bottom of the gap and begins to travel upward slowly; behind it the teeth pivot inward and interlock one by one, left, right, left, right, each landing with a soft tick; the TEETH letters slide off and letters spelling SLIDER slide in.
12-18s: the slider keeps climbing at a steady pace, the alternating interlock ticking on the beat, the joined seam growing neatly behind it.
18-24s: the slider reaches the top and stops; the seam stands fully closed; the SLIDER letters slide off and letters spelling LOCKED slide in and hold; the closed seam gives one small proud pulse.
24-30s: the slider travels back down, the teeth pivoting free one by one in reverse rhythm, and comes to rest halfway with the zipper half open as the LOCKED letters slide away.

Camera: never static, a continuous slow upward drift that climbs with the slider, a gentle push in on the interlocking teeth at the midpoint, and a smooth pull back as the slider descends, with visible parallax between tape, teeth, and slider layers.

Audio: a calm warm narrator, spoken in English, says exactly: "Look closely at a zipper. Two rows of teeth, and one small slider. As the slider travels, it folds each tooth into the next, one by one. Left, right, left, right, a perfect interlock. Down again, and they let go. That's the whole trick, and it never gets old." Diegetic paper sound throughout: a soft tick for every tooth, a dry card scrape as the slider travels, gentle ticks as letters land. Music bed: light pizzicato strings and a small woodblock keeping time with the interlocking teeth.

Use text-to-video at 1080p for 30 seconds, then append both universal blocks.

The story of the last leaf, told in paper on one tabletop set: an old oak of layered charcoal branch pieces at frame centre, its crown dressed in terracotta and ochre leaf pieces, a sage ground line below, and a sky sheet behind that changes with the seasons by physical paper swaps only. One small terracotta leaf on the lowest branch is the character the whole film follows.

0-6s: autumn, thin cream wind-swoosh strips slide across the frame and the crown's leaves lift away one by one, sliding off frame right; the one small leaf stays put, pivoting slightly at its stem joint with each gust.
6-12s: the last other leaves lift away until the branches stand bare except for the one leaf; a stronger wind strip slides straight past it and the leaf pivots hard at its stem but holds on.
12-18s: winter, an indigo sky sheet slides down from the top edge covering the autumn sky, and small chalk-cream snow bits settle one by one onto the branches and the ground line; the little leaf sits still, a snow bit resting on its edge.
18-24s: the thaw, the snow bits lift away one by one, the indigo sheet slides up and off, a warm cream sky settles in, and small sage bud pieces pop open along every branch around the leaf.
24-30s: the little leaf finally lets go, it pivots free of its stem, slides down in gentle zigzag glides, and lands on the ground beside a new sage sprout that unfolds along its crease lines to stand upright next to it.

Camera: never static, it opens on a slow push in toward the crown, drifts closer onto the one leaf through the winter beats, travels down with the leaf's zigzag fall, and settles in a gentle push on the leaf and sprout together at the end, with visible parallax between branch, leaf, and sky layers.

Audio: a warm unhurried narrator, spoken in English, says exactly: "On the old oak, autumn took the leaves one by one. But one small leaf would not let go. Wind pulled. Snow settled. Winter passed. And when the first green buds arrived, the little leaf finally let go, and landed right beside the spring it had been waiting for." Diegetic paper sound throughout: the dry flutter of leaves sliding away, soft ticks as snow bits land, the whisper of the sky sheets swapping. Music bed: a tender solo piano melody, thinning to sparse single notes through winter, joined by warm strings as the buds open and resolving fully as the leaf lands.

Generate at 4K and high quality. Do not append the paper style lock to this photoreal product sheet.

A three-panel product reference sheet composed as one horizontal frame, divided into three equal vertical panels side by side, thin clean separation between panels, the same product rendered consistently across all three.

The product is KRUNCH, a crunchy milk chocolate bar. Its wrapper is uncoated matte paper in a deep coral red, folded tight with a single lengthwise fold seam on the back and short crimped folds closing each end, the segment grid telegraphing through as soft ridges. The bar itself is milk chocolate with a matte-satin finish, moulded four square segments long and two across, each segment gently domed and plain. Inside, the chocolate is dense with toasted puffed rice: any break face shows an open crunchy crumb of chocolate laced with whole crisped-rice pearls and small air pockets, a rough matte fracture, never a smooth solid block.

The brand is KRUNCH, an invented mark: the single word KRUNCH in very heavy, slightly rounded uppercase letterpress capitals, printed in warm cream ink across the centre of the wrapper's front face at nearly full width, visibly debossed into the paper with a faint ink-squash edge. The wrapper carries the wordmark alone and no other text or decoration.

LEFT PANEL, full front elevation, wrapped. The fully wrapped bar photographed square-on from directly in front, upright and centred, its full height and width inside the frame with generous even margin, the KRUNCH wordmark and the ridged segment grid clearly readable.

CENTER PANEL, the open bar. The same bar at the identical scale and position, its coral wrapper peeled open and folded back from one end, the bare chocolate slab slid halfway out of the wrap, and the front pair of segments snapped off, that snapped piece standing beside the bar with its break face turned square to camera so the crunchy interior, chocolate crumb, whole crisped-rice pearls, small air pockets, reads clearly.

RIGHT PANEL, true macro of the break face. The snapped cross-section filling most of the panel at close range: the open crunchy crumb, the crisped-rice pearls seated in the chocolate, the air pockets, the rough matte fracture edge, and the fine grain of the matte-satin chocolate all readable in sharp detail.

18% neutral gray seamless studio backdrop applied uniformly across all three panels, one single flat uniform value corner to corner in every panel, no seam line, no gradient, no hotspot, no vignette, no falloff to black or white, and the identical gray value in all three panels.

Completely flat shadowless illumination in every panel, one enormous soft frontal source at camera position with matched equal fill from camera-left, camera-right, above, and below, so both sides of the product read at exactly the same brightness. No key-and-fill ratio, no modelling, no shadow side, no rim light, no kicker, no specular hotspot. Zero shadow cast onto the background in any panel, no contact shadow, no drop shadow, no ambient occlusion beneath the product. Extremely low contrast, even, milky, catalogue-flat, identical in every panel. Surfaces read at their true material finish: the wrapper matte paper, the chocolate matte-satin, no added gloss, no oily sheen, no blown highlight. The coral-red wrapper, warm cream lettering, and warm brown chocolate render true and consistent across all three panels, identical in value and hue, never washed-out and never cool-shifted by the background. Real material texture, real edge quality, real food detail, visible paper fibre on the wrapper, real fracture texture on the chocolate, never plastic-looking, never a CGI product render. Photographed on a 50mm prime, even sharpness, soft natural film grain. Photographed not generated.

Reference the product sheet, generate at 1080p for 18 seconds and append both universal blocks.

A paper-animation product film: a grey paper city gets its colour back one crunch at a time. The KRUNCH bar comes from the reference sheet and is locked: one size, one proportion, one coral-red shade everywhere; the full wordmark KRUNCH, spelled K-R-U-N-C-H in heavy slightly rounded cream letterpress capitals, stays completely visible and unbroken on the wrapper in every shot where the wrapper appears, never cropped, never covered; the snapped interior is always the same open chocolate crumb laced with crisped-rice pearls and air pockets from the reference, real photographed chocolate, never redrawn as paper; the coral wrapper is the only colour in the opening frame. Everyone and everything else is cut paper: grey card buildings, grey jointed figures with hinged jaws, a grey sky. Colour never fades in, it arrives only as coloured paper pieces physically sliding over grey ones, one clear colour event per beat, never two at once. The frame never freezes, something is moving in every single frame. No narration, no spoken words.

0-2s: the grey street in dull unison, grey figures sliding to work in step, a grey tram gliding through, and one figure at frame centre holding the coral wrapped bar, its full wordmark facing camera, the single point of colour.
2-4s: the figure folds the wrapper back and snaps the bar; the break face turns square to camera and holds one full beat in macro, crumb and rice pearls readable, crumb bits popping.
4-8s: the first bites land on the grid, with each deep crunch exactly one thing gains colour: coral tiles slide over the nearest facade panel by panel, then cream awnings unfold along a shopfront, then the tram is covered in coral panels as it glides past, nearby figures pivoting toward each sound.
8-13s: chocolate squares slide from hand to hand down the street and the figures bite in a wave, one per beat, each bite tiling the next block into terracotta, sage, and cream, window pieces swapping warm one by one, until bunting strips unfold overhead on the last beat of the wave.
13-15s: the full-colour street holds its groove, every figure chewing in rhythm, jaws hinging in ticks, flags flapping on the beat, the tram sliding back through in full coral.
15-18s: the street's two halves slide apart to the left and right edges revealing a chalk-cream brand card where letters K, R, U, N, C, H, the same heavy rounded letterpress capitals as the wrapper, in the same cream, slam in one per beat above the snapped piece, locking on a final crunch.

Camera: never static, a steady lateral track down the street that gently accelerates as the colour spreads, punch-ins on the snap and the first bite, a straight push onto the brand card, visible parallax between building layers throughout.

Audio: no words. The music starts as a sparse muted tick in the grey unison and gains one layer with every crunch, bass, claps, bright plucked strings, every real chocolate crunch and sliding colour tile landing exactly on the drum grid until the full groove carries the finale.

Generate at 4K and high quality with no reference, then append the style lock only.

A single hot-air balloon cut from paper, centred alone on a plain even neutral grey card background, nothing else in the frame. This is a definitive identity sheet for one balloon design: a teardrop envelope built from exactly seven overlapping vertical gore pieces of dyed coral-red stock, every gore the same single coral-red shade, with visible cut lines and a small soft contact shadow where each gore overlaps the next; a narrow chalk-cream throat ring at the envelope's base; four kraft-paper cord strips running straight down from the throat; and a small rectangular basket of woven chalk-cream and kraft paper strips hanging level beneath. The balloon fills most of the frame, upright, seen perfectly flat and face-on, and the grey background stays one even value corner to corner.

Reference the balloon identity, generate at 4K and high quality, then append the style lock only.

The hot-air balloon in this frame is exactly the balloon from the reference image, the same coral-red envelope of seven overlapping gore pieces, the same chalk-cream throat ring, four kraft cord strips, and small woven chalk-cream and kraft basket, identical in design, proportion, and coral shade, reproduced without any change.

That balloon at the moment of launch from a dawn meadow, one wide flat tabletop landscape: the balloon sits low in the left third of the frame, just lifted clear of the ground, above overlapping banks of sage meadow, while a low chalk-cream dawn sun rises between dusty terracotta hills on the horizon and a deep indigo sky still holding the last of the night across the top of the frame.

Every piece in the frame is cut from one fixed set of paper stocks and no others: deep indigo card, dusty terracotta card, warm ochre kraft paper, muted sage-green card, chalk-cream card, and one accent of dyed coral-red stock reserved for the hot-air balloon envelope alone. The indigo reads ink-dark and slightly warm, the terracotta reads dusty and low in saturation, the ochre kraft shows its fibre openly, the sage sits quiet and grey-green, the chalk cream is the lightest value on the table, and the coral red of the balloon is the only saturated note in the frame.

Keep the identity opener and palette paragraphs from the dawn anchor. Replace only its middle scene paragraph with one of these four variants, then append the style lock.

MORNING
That balloon aloft in mid-morning, drifting over a forest, one wide flat tabletop landscape: the balloon rides in the upper-left third of the frame, higher than at launch, above rows of layered sage-green and ochre kraft treetops stacked in overlapping bands filling the lower half, with two small chalk-cream cloud pieces overlapping the balloon's edge and a brightening chalk-cream and ochre morning sky across the top of the frame.

NOON
That balloon at high noon over the mountains, one wide flat tabletop landscape: the balloon rides high in the centre of the frame above jagged overlapping peaks of dusty terracotta and deep indigo card stacked in receding ranges, a small round chalk-cream sun almost directly overhead, and thin warm ochre bands laid across a pale chalk-cream midday sky.

DUSK
That balloon at dusk reaching the coast, one wide flat tabletop landscape: the balloon drifts lower in the right third of the frame, passing tall sea cliffs of layered terracotta and indigo card at frame left, above a wide sea laid in alternating bands of indigo and sage-green card, while a large dusty terracotta dusk sun sits half-folded into the horizon line and warm ochre glow bands cross the darkening sky.

NIGHT
That balloon landing at night, one wide flat tabletop landscape: the balloon settles low in the centre of the frame, its basket just touching down in a dark meadow of near-black sage banks, under a deep indigo night sky filling the upper frame, with one round chalk-cream moon piece and small torn chalk-cream star bits scattered across it, the coral envelope reading as the warmest and brightest thing on the table.

Reference the identity sheet first, then the five anchors in order. Generate at 1080p for 30 seconds and append both universal blocks.

One coral-red hot-air balloon crosses one full day in one continuous tabletop world. The first reference image is the balloon identity sheet: that exact balloon, coral envelope of seven overlapping gore pieces, chalk-cream throat ring, four kraft cords, woven chalk-cream and kraft basket, is the only balloon in the film and never changes design, proportion, or shade. The remaining five references are the five successive states of the scene in strict order: dawn launch over the sage meadow, morning over the layered forest, noon over the mountain ranges, dusk at the sea cliffs, night landing in the dark meadow. The film settles fully into each state's exact layout before the next transition begins; two states never blend or average.

The transitions are the showpiece: at each change the outgoing landscape physically comes apart layer by layer, foreground first, pieces folding closed along crease lines, sliding out past the frame edges, or being covered, and the incoming state reforms piece by piece in the same order, while the balloon alone never leaves the frame, drifting to its next position and carrying the continuity.

0-6s dawn: the balloon lifts from the meadow as the cream sun rises between the terracotta hills. Transition on the downbeat: the meadow banks slide down and out in one long dry scrape while the forest bands slide up and snap into place.
6-12s morning: the balloon drifts higher past two cream clouds. Transition: the tree bands fold flat with soft snaps while indigo and terracotta peaks hinge up from the horizon line.
12-18s noon: the balloon rides high over the ranges under the small cream sun. Transition: the peaks part and slide off both sides with a scrape as banded sea strips slide in from frame right and cliff stacks build at frame left.
18-24s dusk: the balloon eases lower past the cliffs while the terracotta sun folds into the horizon. Transition: a deep indigo night sheet slides down with a heavy soft settle as torn star bits and the cream moon piece tick into place one by one.
24-30s night: the balloon descends, its basket touches the dark meadow, and the envelope gives one final gentle settle on the last beat.

Camera: never static, one continuous slow lateral drift left to right across the entire thirty seconds, as though the table were one long unrolled journey, with a gentle push in on the final touchdown and visible parallax between the stacked layers throughout.

Audio: no spoken words. One continuous music bed at one steady tempo, each state lasting one even musical phrase, warm acoustic guitar, soft hand percussion, and a light string pad, and every transition lands exactly on a downbeat, marked by its own signature paper accent: the long dry slide of the meadow leaving, the soft snaps of the forest folding, the scrape of the parting peaks, the heavy soft settle of the night sheet, and a quiet final tick as the basket touches down.

Upload your portrait as the GPT Image 2 reference. Generate at 4K and high quality. Paste part 2 immediately after this text, then append the style lock.

A three-panel character reference sheet composed as one horizontal frame, divided into three equal vertical panels side by side, thin clean separation between panels, the same paper cutout puppet rendered identically across all three.

The puppet is a man in his late twenties of medium athletic build and even proportion, drawn and printed in a gently cartoonish hand-illustrated style, the likeness of the uploaded person simplified into flat printed ink shapes, never photographic. The face reads like a hand-pulled screen print on card: flat ink layers with a slight registration offset, visible paper grain through every printed area, no photographic skin texture, no photo-real rendering anywhere. His skin tone is one flat warm ink tone printed evenly onto the head, neck, and hand pieces. His hair is thick and dark brown, swept up and back off the forehead, cut as three overlapping layered paper shapes with scissor-cut edges. His printed features are simplified and friendly: oval warm-brown eyes, bold single-stroke dark eyebrows, a straight nose drawn as one simple printed line, a wide easy smile with simply drawn teeth, and short stubble suggested by a flat printed tone patch along the jaw, chin, and upper lip.

The head is one flat piece of medium-weight card with the illustrated face printed onto it; the hair layers, torso, upper arms, forearms, open hands, hips, thighs, shins, and shoes are each separate pieces of the same matte stock family. The clothing sits on top as separate layered paper pieces: a black casual zip jacket cut from matte black card as front panels, collar, and sleeve pieces with a faint printed fabric grain, over a chalk-white crew-neck tee piece, with dark indigo trouser pieces and simple off-white shoe pieces. Round brass-coloured paper fasteners sit visibly at both shoulders, both elbows, both hips, and both knees, and every layer casts a small soft contact shadow onto the layer behind it.

Paste this directly after part 1 in the same GPT Image 2 request, then append the style lock.

LEFT PANEL, full body front view, head attached. The full paper puppet stands squared to camera, framed from the top of the head down to the feet, arms relaxed at the sides, hands open and loose, weight even across both feet. The head is present, whole, and facing camera, cut from paper as a single flat piece with the face printed onto it, the same stock and the same cut quality as the rest of the figure. Full headroom above the head. Nothing is removed, hollowed, cut away, or terminated anywhere on the figure.

CENTER PANEL, full body rear view, head attached. The same puppet photographed from directly behind, standing straight, arms relaxed at the sides, hands loose, weight even across both feet, from the top of the head down to the feet at the identical scale and position as the left panel. The back of the head piece, the back faces of the clothing layers, the rear side of every paper fastener, and the raw reverse of the stock all readable from behind.

RIGHT PANEL, tight chest-up view, identity lock. The same puppet framed from just above the top of the head down to the collarbones and the very top of the clothing layer only, the head filling most of the panel, a true close-up. Squared to camera, head level, facing directly forward. The printed facial features, the cut edges of the hair pieces, the paper grain of the head stock, and the fibre lift along every cut edge all clearly readable at close range.

The paper stock, the cut edge quality, the fibre texture, and every colour render identical in value and hue across all three panels, and the tabletop surface behind the puppet is the same even neutral surface at the same brightness in every panel.

Reference the completed puppet sheet, generate at 1080p for 15 seconds and append both universal blocks.

The jointed paper puppet from the reference sheet, the reference shows the same puppet front, back, and in close-up; his gently cartoonish printed face, swept dark-brown layered paper hair, black paper zip jacket over a chalk-white tee, dark indigo trousers, off-white shoes, and round brass paper fasteners at shoulders, elbows, hips and knees all carry over exactly. His face is a flat printed illustration on the head piece and never turns photographic. His expression changes across the film only by one flat printed face piece being swapped for another in a quick physical slide, the head shape, hair pieces, and skin tone never change. The set is a tabletop basketball half-court: an ochre kraft court line along the bottom of the frame, a terracotta hoop with a chalk-cream backboard and a paper-strip net on an indigo pole at frame right, the puppet standing upright at frame left.

0-5s focus: the puppet dribbles a small ochre paper ball, his right forearm hinges at the elbow fastener while the ball slides straight down to the court line and back up to his flat open hand, three even bounces, wearing his focus face: brows printed low and level, eyes narrowed, mouth a small flat line.
5-10s anticipation: both arms pivot up at the shoulder fasteners and the ball slides away from his hands along one high clean arc toward the hoop; the instant it leaves his hands his face piece swaps with a quick slide to anticipation: wide round eyes, brows printed high, a small open o mouth; the ball meets the rim and pivots flat around the rim's centre in one full slow suspenseful circle while he holds frozen mid-reach.
10-15s joy: the ball drops through, the paper-strip net splays outward as it passes, torn paper confetti bits pop up and scatter, and his face piece swaps again to joy: eyes curved shut in happy crescents and a huge open grin, while his arms pivot straight up in victory and his feet slide a small celebratory shuffle on the court line.

Camera: never static, it opens on the puppet, drifts laterally with the ball's arc toward the hoop, then arcs gently back to hold both puppet and hoop in frame for the drop and the celebration, visible parallax between the stacked layers throughout.

Audio: no spoken words. Diegetic paper sound: a soft card tap on every dribble, a dry scrape as the ball circles the rim, a papery swish through the net, a crisp tick on each face-piece swap, tiny flutters from the falling confetti. Music bed: a bouncy bass line, tight handclaps and a bright melodic whistle at an upbeat tempo, holding one suspended beat while the ball circles the rim, then landing back on the groove for the celebration.