Seedance 2.5 Prompts
25 tested prompts, each paired with the exact clip it generated. Copy any prompt or run it on Astorie. Single-pass takes up to 30s; every clip generated in-house.
Prompt recipes
The Masquerade Opens
Cinematic · Musical — 16:9, 15s (shot C-13).
Style: Hollywood musical grand-scale, a European court masquerade ball; baroque palette of candle-gold and deep crimson; high-saturation dramatic lighting; alternating slow-motion and real time; cinematic. Duration: 15s. Scene: A Venetian-style palace ballroom — a marble double staircase, three enormous crystal chandeliers overhead, gilded wall sconces, tall windows onto the night, a dance floor polished to a mirror. Cast: A Black male master of ceremonies in a gold half-mask (mouth and jaw exposed), a black-and-gold embroidered tailcoat, holding a cane; twenty masked couples on the floor, each mask distinct (feathers, bird beak, cat face, sun-and-moon), gowns matched one-to-one to each mask. [00:00-00:05] The ballroom is dim, only scattered candlelight. The MC stands alone at the top of the marble stair, taps the step twice with his cane, and on the downbeat sings the first line: "Welcome to the masquerade!" — on the exact syllable, all three chandeliers and every wall sconce blaze up at once; the twenty couples are already in position and sweep into a waltz on the same beat. [00:05-00:11] One continuous shot dives from the stair-top down into the floor, threading between spinning dancers: skirts open into whirling circles, feathers and gold dust glitter in the candlelight, two lines of dancers part into a corridor, the MC walks and sings along it, gold flakes drifting down from above, chandelier reflections rippling across the floor. [00:11-00:15] On the final accent every dancer freezes in the same frame — bowing toward camera, masks forward — the MC at center throwing his arms wide; all candlelight snaps down to a single chandelier spotlighting him; the camera pulls back fast over the ballroom into a wide of the entire palace hall. Audio: A string waltz laced with dramatic timpani; the English lyric lip-synced; skirt rustle and footfalls locked to the 3/4 beat; the whoosh of every candle igniting on the downbeat. Constraints: The MC's half-mask must keep the mouth visible with accurate lip-sync; each dancer's mask and gown stay matched and never swap or deform; waltz spins stay locked to 3/4; on the light cue the dancers are already placed, not scrambling; the freeze lands all dancers on one frame with the light-drop on the same accent; gold flakes fall with natural physics.
Recommended model:seedance-2
Leviathan at the Shore
Cinematic · Sci-Fi — 16:9, 15s (shot C-03).
Style: Large-format disaster film, photoreal, storm daylight, high dynamic range, 8K. Duration: 15s, three shots. Scene: A coastal city on the night of a typhoon — torrential rain, a heaving sea, the lights of a sea-crossing bridge blooming in the downpour. [00:00-00:05] Shot 1 | high aerial: a vast dark mass rises slowly from the sea, seawater sheeting off its surface like waterfalls; the sea-crossing bridge beside it reads as tiny, giving the scale. [00:05-00:10] Shot 2 | ground-level low angle: a tentacle thicker than a building slams down through the rain, wraps an office tower and slowly tightens; the glass curtain wall shatters panel by panel, shards spinning down in slow motion. [00:10-00:15] Shot 3 | street-level handheld: a crowd flees with their backs to camera, standing water flying up, the frame shaking hard. The camera finally tilts up — a single enormous vertical pupil opens slowly in the rain and contracts to focus. Audio: Storm and heavy rain, a sub-bass infrasonic roar, breaking glass, screaming crowd; no score. Constraints: The creature's proportions stay consistent throughout; the tentacle carries real weight and inertia; rainwater flows over its surface at the correct scale.
Recommended model:seedance-2
The Frozen Crossroads
Cinematic · Sci-Fi — 16:9, 15s (shot C-02).
Style: Urban sci-fi, cinematic realism, clashing warm-cool neon, shallow depth of field, 35mm. Duration: 15s. Scene: A busy intersection at night, dense foot and vehicle traffic, wet reflective ground after rain. [00:00-00:05] Medium shot: a Black woman in a black trench coat stands at the curb, sharp jaw-length bob, raising a hand to touch a metal band on her left wrist. Everything around her flows normally. [00:05-00:10] The instant she presses the band, everything except her freezes completely — pedestrians locked mid-stride, raindrops suspended in air, headlight beams frozen. Only she can move. The camera circles her slowly 180°, passing through the still crowd and the hanging rain. [00:10-00:15] She slips an envelope from a frozen passer-by's hand, steps back into place, and touches the band again. The world resumes at once, raindrops fall, and the passer-by looks down, baffled, at his suddenly empty hand. Camera: The orbit holds a constant speed; during the freeze only she and the camera move. Audio: At the freeze, ambient sound drops away to a high-frequency ringing; on resume every sound rushes back at once. Constraints: During the freeze all other people and raindrops are completely still with no micro-motion; after resume, motion continues seamlessly from where it stopped.
Recommended model:seedance-2
Deep-Space Waking
Cinematic · Sci-Fi — 16:9, 30s (shot C-01).
Style: Hard sci-fi cinematic, cold teal-blue, practical-set feel, no plasticky CG, 4K. Duration: 30s, single continuous shot. Scene: A spaceship hibernation bay, cramped, dense with piping, the only light from cabin indicator LEDs and starlight through the porthole. Cast: A young Latina in a hibernation pod — dark brown curly hair, skin drained of color from the long sleep, a dark grey cryo-suit; face and hair consistent throughout, no warping or drift. [00:00-00:08] The camera pushes slowly forward; in the dark only scattered indicator lights blink in rhythm, gas glowing faintly inside the pipes, advancing toward the one pod still lit. [00:08-00:16] The pod vents — white cold fog jets from the seams and sinks across the floor. The lid rises mechanically. Push to a facial close-up: eyelashes tremble, eyes open slowly, pupils dilate then contract to focus. [00:16-00:24] She lifts an arm to brace against the pod wall and sit up, the arm trembling weak from long sleep, fingertips leaving condensation streaks on the metal. She coughs twice, hard, breathing ragged; the camera drops and pulls back with her motion. [00:24-00:30] She sits up and turns to the porthole; the camera pans along her line of sight — beyond the glass a huge ringed planet turns slowly, its reflected light lighting half her face, her expression shifting from dazed to disbelief. Audio: Ship low-frequency hum, venting hiss, the mechanical lid, coughing and ragged breath, a synth drone swelling; no dialogue. Constraints: 30s of continuous time with no jumps; correct finger count; the light logic transitions naturally from cabin LEDs to the planet's reflected glow.
Recommended model:seedance-2
End of the Corridor
Cinematic · Thriller — 16:9, 15s (shot C-11).
Style: Psychological horror, very low light, the only sources a flashlight and an emergency light at the far end; cold green and dark red; high-ISO grain. Duration: 15s. Scene: The long corridor of an abandoned hospital — peeling walls, scattered medical files and broken glass, a flickering emergency light on the ceiling. [00:00-00:05] First-person flashlight POV; the beam sweeps slowly across the dark, catching wall stains, fallen ceiling fragments, a half-open door. Breathing is quick but controlled. [00:05-00:09] As the beam reaches the far end, it briefly catches a blurred upright human silhouette in the deepest dark — only half a second, then the shaking hand sweeps the beam off it. The breath stops. The beam trembles back — empty. [00:09-00:13] A metal clang from behind; the view whips around, beam thrashing over empty ground. The breathing settles; the room falls almost totally silent but for the emergency light's electric buzz. The beam lowers to the floor — a line of wet footprints leads straight toward camera, ending half a meter away. The beam lands on the end of the prints: a pair of pale bare feet, ankles vanishing up into the dark. [00:13-00:15] The beam jerks up the soaked hospital gown — at 13.5s a chalk-white face lit from below fills the beam an arm's length away, sharp light and shadow, eyes wide. It lunges at camera, the face blowing out in overexposure as it rushes in, a shrill sting hits, the frame jolts, and 0.5s later it cuts hard to black, held to the end. Audio: Suppressed breathing, glass crunching, distant dripping, a metal clang, the emergency light's buzz; near silence from 11–13.5s; on the 13.5s scare a shrill sting and low-frequency hit burst together; after the cut to black all sound stops dead. No score. Constraints: Nothing is lit in the dark without a light source; the distant figure appears only for the specified half-second and stays unclear; the face is inside the beam, lit bottom-up, with clear light/shadow; the reveal runs feet → gown → face continuously without jumping; the lunge is visible fast motion, not a teleport; from face to black-out is under one second; the cut to black is clean with no fade.
Recommended model:seedance-2
The Driveway Goodbye
Cinematic · Drama — 16:9, 15s (shot C-10).
Style: American family drama, warm film tone, low-angle backlight at dawn, slight handheld, autumn palette. Duration: 15s. Scene: The driveway of a suburban house; a station wagon with its trunk crammed full of suitcases and boxes; yellow leaves falling from the street trees. Cast: An 18-year-old daughter — chestnut long hair, a college hoodie; a father in his 50s — plaid shirt, greying temples. [00:00-00:06] The daughter shoves the last box into the trunk and pushes it firmly shut. The father stands at the far side of the driveway, hands in his pockets, opens his mouth and says nothing. Dawn light rakes across the space between them. [00:06-00:11] She walks over and hugs him. He pats her back at first, then his arms tighten suddenly, holding a couple of seconds longer, chin on the top of her head, eyes closing once. She lets go, smiles, gets in the car. [00:11-00:15] The car rolls slowly out of the driveway; the daughter watches him in the rear-view mirror. The father stands alone, giving a small wave, and doesn't lower his hand even after the car leaves frame. It ends on him and the empty driveway. Audio: Dawn birdsong, the muffled thud of the trunk, the engine starting, tires over dry leaves; a minimal piano enters only at the end; no dialogue. Constraints: Both faces and outfits stay consistent throughout; the two beats of the hug (back-pat → tighten) are clearly distinct; the father in the mirror and the father on the driveway stay positionally consistent; no subtitles.
Recommended model:seedance-2
The Sinking Chair
Cinematic · Comedy — 16:9, 12s (shot C-14).
Style: American mockumentary office comedy, single-camera handheld zoom, flat fluorescent office, deadpan humor, realistic. Duration: 12s. Scene: A glass-partitioned conference room; four colleagues around a long table; charts taped all over the whiteboard. Cast: The lead is a white man in his 30s, shirt and tie, overconfident; the other three colleagues are of varied ethnicities. [00:00-00:05] The lead finishes his last slide standing, says smugly, "The numbers speak for themselves," and drops back into his chair with a flourish, hands laced behind his head, leaning back. The handheld camera drifts slightly — documentary feel. [00:05-00:09] The pneumatic chair lets out a hiss of escaping air and begins to sink, extremely slowly. He notices but pretends nothing is wrong, holding the hands-behind-head pose as he descends, eyes starting to dart. The colleague opposite shakes trying not to laugh; the camera snap-zooms to the lead's face. [00:09-00:12] The chair bottoms out; only his head remains above the table. A full second of silence. Deadpan, he turns his head straight to camera and blinks once. Freeze. Audio: The hiss of the pneumatic chair, the colleagues' stifled laughter, the HVAC hum; no score; English lines lip-synced. Constraints: Keep the mockumentary handheld feel and the snap-zoom; the chair's descent must be slow and continuous, never a jump-cut; the lead holds the hands-behind-head pose the whole time; the final look to camera is the key comedic beat.
Recommended model:seedance-2
Backyard Cartoon Dog
Animation · Retro Cartoon — 16:9, 10s (shot A-01).
Style: 1990s American Saturday-morning TV cartoon, thick black outlines, bright flat color fills, exaggerated squash-and-stretch, limited-animation frame feel, cheerful and retro. Duration: 10s. Scene: A suburban American backyard, a white picket fence, a green lawn, a smoking grill with a sizzling sausage on the grate. Cast: A lanky beagle-type cartoon dog — white body with big brown patches, two long brown floppy ears hanging by its face, a long snout, a round black nose, a thin tail, big eyes, exaggerated expressions. [00:00-00:03] The dog stares at the sausage on the grill, its eyes ballooning to the size of plates, drool dripping, one floppy ear springing half-up then falling. It tiptoes closer in a classic cartoon sneak, each step tapping down lightly, ears bouncing with the gait. [00:03-00:07] It pounces at the sausage — the grill lid snaps up and flings it skyward, its lanky body stretching into a long line, ears streaming back like ribbons, and it smacks into the picket fence, flattening into a pancake that slowly slides down onto the grass, then pops back into shape with a boing. [00:07-00:10] Dizzy, it sits up with a ring of little stars spinning over its head, its ears tied in a knot that unties itself. The sausage falls from the sky and lands squarely in its open snout. It blinks, gulps it down in one bite, thin tail wagging like a fan. Freeze. Audio: Classic cartoon SFX — slide whistle, spring boing, cymbal crash on impact; a cheerful small-jazz-band score. Constraints: The character is a dog — the long snout and long floppy ears are clearly visible throughout, never a cat's short round face; squash-and-stretch can be extreme but the coat color and patch positions stay consistent; after flattening it returns to shape; the head-stars appear only in the specified beat; keep the 2D flat-color look, never 3D.
Recommended model:seedance-2
Rooftop Parkour
Animation · Comic — 16:9, 12s (shot A-06).
Style: Modern American comic-book film animation with retro print texture — halftone dots for shadow, CMYK misregistration (magenta and cyan edges slightly split at contours), the character animated on 12 frames-per-second with a staccato snap while the camera and background stay buttery smooth, the frame occasionally splitting into comic panels, high-saturation graffiti-neon palette. Duration: 12s. Scene: Old-town rooftops of a big American city at dusk — wooden water towers, antennas, clotheslines; a steel bridge and an orange-purple sky in the distance. Cast: A Black teen traceur in an oversized fluorescent-green hoodie, a backpack full of spray cans, wearing over-ear headphones. [00:00-00:04] The teen sprints off the rooftop edge — his motion snapping on 12s while the camera pan and drifting clouds stay perfectly smooth, both frame rates coexisting in one image. On the takeoff the frame splits into three comic panels — three angles of the same move — then merges back to one. [00:04-00:08] He leaps a gap between buildings, slow motion: a halftone-dot motion trail streaks behind him, the magenta/cyan misregistration at his contour intensifies; on the impact of the landing roll a graffiti-style onomatopoeia "WHAM!" bursts out, expands with a shockwave, and dissipates. [00:08-00:12] He slides down a fire-escape rail onto a platform, pulls off the headphones and looks back at the skyline in the sunset; the camera orbits him, the halftone dots in his facial shadow clearly visible. It freezes on a comic-cover composition with hand-drawn speed-line borders around the frame. Audio: A hip-hop beat, the clatter of spray cans in the backpack, a bass hit on the landing synced to "WHAM!", a distant city siren. Constraints: The character's 12fps staccato and the smooth background must coexist (this is the soul of the style); halftone dots and color misregistration stay consistent in strength throughout; the onomatopoeia appears once only, spelled correctly as WHAM!, with the impact then dissipates; the three panels show the same move from different angles; hoodie color and backpack stay consistent.
Recommended model:seedance-2
English Countryside
Animation · Watercolour — 16:9, 15s (shot A-02).
Style: European children's-book watercolour animation, soft bleeding washes, visible paper texture and pigment pooling at the edges, characters drawn as line art with light color, bright and luminous, English pastoral. Duration: 15s, single shot. Scene: The English countryside in late summer — a narrow gravel lane through golden wheat fields, low drystone walls on either side, sheep scattered on a distant hillside under a big oak, a stone arch bridge spanning the valley. [00:00-00:05] Wide establishing shot; the wind blows right to left, the whole wheat field undulating like water, cloud shadows drifting slowly over the hillside, a little steam train puffing white smoke across the far stone bridge. [00:05-00:10] The camera lowers to road height; a red-haired freckled girl in a blue pinafore cycles in from deep in the frame, a bunch of wildflowers in the basket, her skirt lifting in the wind, the sheep by the wall looking up at her. [00:10-00:15] The camera pans to follow her past. She looks up at the sky and a flock of white doves startles up from the oak by the field and flies away. She exits frame right; it holds on the empty gravel lane and the swaying wheat. Audio: Wind through the wheat, a bicycle chain, dove wingbeats, a far train whistle, a minimal single-note piano. Constraints: The watercolour brushwork stays stable and doesn't drift; the character design and clothing colors stay consistent; single shot, no cuts.
Recommended model:seedance-2
A Rose in Black & White
Animation · Noir Comic — 16:9, 12s (shot A-04).
Style: Animated American noir comic (hard-boiled graphic-novel style), high-contrast pure black-and-white, coarse ink hatching and halftone dots for greys, large silhouette ink masses, hard-boiled noir mood. Duration: 12s. Scene: A rain-soaked city rooftop at night, downpour, the background black silhouettes of buildings with water towers and fire escapes, distant skyscraper lights as scattered white dots. Prop: He holds a single red rose — the only color in the whole piece. [00:00-00:04] Extreme low angle; a man in a long trench coat and wide-brim hat stands with his back to camera at the roof edge, the coat's hem whipping violently in the wind, rain slashing the frame in dense diagonal white lines, the red rose in his lowered hand jumping out sharply against the pure black-and-white. [00:04-00:08] Cut to a facial close-up, the hat brim's shadow covering the upper face, rainwater running off the jawline; he slowly lifts his head, his eyes two white dots in the hatched shadow, the cheekbone transition rendered in halftone dots, jaw clenched. [00:08-00:12] Pull to a medium shot; he opens his fingers — the red rose drops, the camera following it straight down past layer after layer of hatched buildings, the rose finally landing in the wet standing water of the street, red rippling out across the black-and-white surface. Freeze. Audio: Dense rain, muffled thunder, the coat tearing in the wind, a soft splash as the rose hits the water; no dialogue, no score. Constraints: Strictly black-and-white except the rose's red; hatching and halftone density stay stable; the rose is the same single flower from hand to water, shape unchanged; the fall obeys gravity; hat and coat design stay consistent.
Recommended model:seedance-2
Shrine Blossom Duel
Animation · Sakuga — 16:9, 12s (shot A-07).
Style: High-level Japanese TV-anime battle animation (sakuga), cel shading, sharp clean linework, smear frames for high-speed action, black-and-white inverted impact frames at each clash, speed lines, 2D hand-drawn texture. Duration: 12s. Scene: The long stone stairway of a shrine at dusk, cherry trees in full bloom on both sides, petals blowing sideways in the wind, a torii gate silhouetted backlit at the top of the steps, an orange-purple gradient sky. Cast: A white-haired, red-eyed teenage swordsman — a black stand-collar school uniform, holding a wooden sword; his opponent a grey-robed figure in a white fox mask, a short blade hidden in the sleeve. Both keep their outfits and silhouettes consistent throughout. [00:00-00:04] Standoff: the boy at the bottom of the steps, the fox-masked figure at the top; a gust sweeps a mass of cherry petals across the frame. The boy slowly raises the wooden sword before him, the fox-mask's short blade sliding silently from the sleeve. Only the petals move; the two stay still, tension built by composition and wind. [00:04-00:08] Clash: both push off and sprint at once, petals on the steps kicked up by the airflow. At the instant they cross, 1–2 frames of black-and-white inverted impact frames cut in, sparks bursting where the blades meet, high-speed swings shown with smear frames, the background stretched into speed lines, the sword-pressure blasting the surrounding petals into a ring. [00:08-00:12] Resolve: the two pass each other and land back-to-back at the same moment, holding a beat. A crack splits the fox mask and half of it slides off, revealing half a smiling face; the boy sheaths his sword with a crisp clack. The petals drift slowly back down onto the steps. Freeze on the two silhouettes backlit under the torii. Audio: Wind and rustling petals, the burst of the push-off, the metallic ring of the clash landing on a taiko hit synced to the impact frame, the crack of the splitting mask, the sheathing sound; sparse shakuhachi and shamisen accents. Constraints: The 2D cel look must never become 3D; impact frames appear only at the clash for 1–2 frames, black-and-white inverted; smear frames only for the high-speed swings; cherry petals move coherently with the sword-pressure and airflow; both fighters' outfits and hair stay consistent; the fox-mask figure's face must not show before the mask cracks.
Recommended model:seedance-2
Couture Runway
Fashion · Runway — 16:9, 15s (shot F-01).
Style: Paris haute-couture live-show footage, a black mirror runway, a single top spotlight, the audience sunk in darkness as silhouettes with scattered phone screens, cinematic photography. Duration: 15s. Cast: A tall Black woman model, shaved head, in a jewel-blue couture gown — a rigid sculptural bodice, the skirt dozens of layers of fine tulle. [00:00-00:06] Low front angle; she walks from deep in the runway toward camera, her pace steady on the beat, the spotlight tracking her, the tulle skirt trailing a delayed wave with each step, the mirror floor reflecting a full inverted image. [00:06-00:11] She stops at the runway's front, plants her stance, one hand on hip, chin slightly raised. Cut to a side view slowly orbiting a quarter turn, showing the gown's sculptural cut, the tulle half-transparent in the spotlight. [00:11-00:15] She turns to walk back, the skirt swinging out in a full arc then falling in layers. A male East-Asian model in a black suit passes her at mid-runway, their paces matched, not interfering. It holds on her receding back and the mirror reflection. Audio: Low-frequency show electronica, the crisp click of heels on the mirror floor, a burst of shutter clicks; no dialogue. Constraints: The gown's color, layer count and cut stay consistent throughout; the pace stays steady; the two models' outfits never swap; the mirror reflection stays in sync with the motion.
Recommended model:seedance-2
Haute Couture Quick-Change
Fashion · Quick-Change — 9:16, 12s (shot F-02).
Style: High-fashion quick-change short, a dark studio with a single runway spotlight, the light color switching with each collection, high-fashion editorial feel, viral vertical pacing. Duration: 12s, 9:16 vertical. Cast: A young Slavic woman — high cheekbones, grey-blue eyes, long straight ash-blonde hair, tall. The same face and identity throughout; hair and makeup change with each look. [00:00-00:03] Look 1 | warm-white spotlight | French maritime couture: a black-and-cream horizontal-striped long-sleeve top cinched at the waist over a tiered black-and-white feather-fringe skirt, a little white sailor cap. Hand on hip, feathers trembling as she shifts her weight; chin up to camera; a fast spin in place — hair and feathers fly up and blot out the frame — [00:03-00:06] Look 2 | magenta spotlight | surrealist couture: as the hair falls she's already in a shocking-pink structured strong-shouldered mini blazer-dress, gold relief buttons, a pair of gold sun-relief earrings, a gold keyhole belt, hair in a sleek bun. She frames her face with both hands in a surreal gesture, tilts her head to camera, the lens pushing in on the downbeat. She spins again — [00:06-00:09] Look 3 | gold warm light | Italian glamour: now a bright-yellow satin one-shoulder gown, high slit, gold chain necklace, big blonde waves loose, vivid makeup. She steps forward, the slit falling open, satin flowing like liquid in the gold light, and glances back with a smile. One last spin — [00:09-00:12] Look 4 | cold blue hard light | sculptural futurism: a jet-black high-gloss body-con dress — exaggerated sculptural shoulders, an extreme cinched waist — wet-look hair slicked back, sharp eye makeup. She opens both shoulders into a power stance and freezes, hard light drawing a sharp highlight down the fabric. End: a rapid recap of all four looks, one frame each. Audio: A driving electronic beat over strings, each change landing on the downbeat; feather rustle, earring chime, satin slide, stiff-fabric friction switching with each collection. Constraints: Face and identity stay strictly unchanged — only clothing, hair/makeup and lighting change; the change happens only in the instant the spin hides her behind her hair, with no ghosting or dissolve; the four silhouettes and colors differ enormously (feathers / hot-pink-gold / yellow satin / black sculptural) each with correct fabric physics — feathers tremble, satin drapes and flows, the hard shoulder line doesn't wobble; each look is clearly readable for at least 2 seconds; no logos or text.
Recommended model:seedance-2
Underpass Editorial
Fashion · Editorial — 16:9, 12s (shot F-03).
Style: Fashion-magazine editorial, alternating fisheye and 35mm, an urban pedestrian underpass, green fluorescent tubes and strobing flash, film grain, high-end grit. Duration: 12s. Cast: A Black male model — locs tied up, an oversized coat and leather boots; a white female model — platinum pixie cut, a silver metallic jumpsuit. Their outfits never swap. [00:00-00:04] Fisheye wide, the two stand back-to-back in the center of the underpass, each looking the opposite way; the camera orbits them fast, the wide-angle distortion bowing the tube lights into arcs. [00:04-00:08] Cut to 35mm; a strobe-flash rhythm starts: on each flash the two snap into a new pose, four times total, poses crisp like burst stills. Between flashes the frame drops back to fluorescent-green shadow. [00:08-00:12] The camera pushes suddenly into garment macro: the coat's tweed weave, the reflective creases of the jumpsuit's metallic fabric, the zipper on the side of a boot, about a second each. Finally it pulls back to both models full-length, both looking to camera at once. Freeze. Audio: The electronic charge-and-release of the flash, the fluorescent tubes' hum, a cold electronic track; no dialogue. Constraints: The two models' outfits, hair and poses never swap; each strobe pose is crisp with no motion blur; the macro fabric details match the garments seen in the full-length shots.
Recommended model:seedance-2
Baroque Pearl Editorial
Product · Jewelry — 3:4, 10s (shot F-05).
Style: Modern affordable-luxury jewelry editorial, studio with a light-grey seamless backdrop, soft light plus one side hard light, extreme-shallow macro, real skin detail, magazine quality, 8K. Duration: 10s, hard cuts between the three shots. Cast: A deep-skinned Black model in a black T-shirt, only body details shown (ear, wrist, lower half of the face and collarbone) — never a full face. Subject: Gold jewelry with irregular baroque pearls — every pearl a lopsided freeform (teardrop, knobbly, flat-round, all different); design language: leaf-shaped openwork gold, half-pearl-half-chain joins, satellite chains studded with tiny gold beads, star charms; no brand text. [00:00-00:04] Shot 1 | ear macro: a gold ear-cuff, its placement clear — two openwork leaf-shaped gold plates overlapping to clamp firmly onto the mid-cartilage of the outer ear, the gold hugging the ear's curve with a real contact shadow where it meets skin; a line of tiny pavé along the veins; from the leaf tip a very fine ~3cm gold chain hangs naturally in front of the lobe, ending in a lopsided teardrop baroque pearl below the lobe. She turns her head very slightly — only the dangling chain and pearl sway a little and settle, the cuff itself dead still; the side hard light runs irregular highlights over the pearl's bumps and casts fine shadows through the openwork onto the skin; a soft rim light on the ear's edge. [00:04-00:07] Shot 2 | hard cut to wrist macro: the wrist rests against the black T-shirt, two bracelets stacked — the outer one a "half-pearl-half-chain": one half strung mini freeform pearls, the other half a gold twist chain, joined by a sculptural hook clasp; the inner one a fine satellite chain of tiny gold beads carrying three charms — a knobbly baroque pearl, a four-point star plate, a flat-round pearl. The wrist turns slowly outward, the charms knocking and swinging, each pearl's distinct shape clearly readable, every link of the twist chain sliding, the hook clasp flashing a line under the hard light. [00:07-00:10] Shot 3 | hard cut to lower face and collarbone: two gold satellite chains of different lengths stacked — a short one hugging the throat in a collar arc, a long one draping two fingers lower; both come around from the back of the neck, both sides of the neck always clearly showing two strands. On the short chain five or six small pearls of varied size scatter asymmetrically (denser left, sparser right); at the center of the long chain hangs a thumbnail-sized freeform teardrop baroque pearl, capped at the top with a simple gold bail. She shows a faint smile (only the corner of the mouth and jawline visible), the pearl surfaces reflecting soft light and the warm reflection of skin, the small beads rising and falling gently with breath at the collarbone. Freeze. Audio: The fine metallic click of charms, fabric friction, quiet room tone, a light lazy string piece. Constraints: The ear-cuff stays fixed on the cartilage the whole time — its contact point does not shift, float, or detach; only the dangling chain sways on the head turn; on the collarbone shot the two stacked chains stay independent, never tangling or merging, two strands visible on each side at all times; every baroque pearl is irregular and different, each one's shape never changing into a regular round bead; pearls and metal never fuse or deform; the leaf openwork, star charm and hook clasp stay structurally clear; movement obeys gravity and inertia; skin is real with pores and texture; hard cuts between shots, no pans; jewelry hugs skin without clipping; no brand text.
Recommended model:seedance-2
Perfume, Minimal
Product · Fragrance — 16:9, 10s (shot P-01).
Style: High-end perfume ad, minimal composition, broad diffuse light plus one rim light, seamless off-white background, shallow-DOF macro, 4K. Duration: 10s, single shot. Subject: A square clear-glass perfume bottle, heavy body, the liquid inside a pale amber, a gold cylindrical cap, a single hairline silver line on the front, no text. [00:00-00:04] Front macro, the camera panning extremely slowly, light sweeping across the bottle's edges, the glass thickness refracting the liquid into layers, the surface swaying very slightly. [00:04-00:07] The camera rises and pulls back slowly, revealing the full silhouette and its reflection on the matte surface, the gold cap throwing a thin elongated highlight. [00:07-00:10] A faint wisp of white mist drifts down past the bottle from above, light passing through it into a soft beam, the bottle briefly glowing within it, then the mist clears and the frame returns to pure and clean. Audio: A soft ambient pad; no dialogue. Constraints: The bottle's proportion, color and cap material stay consistent; glass refraction and reflection are physically correct; the camera moves at constant speed with no shake; no text or logo, no plasticky look, no fingerprints, no hands in frame.
Recommended model:seedance-2
Pour-Over
Product · Beverage — 16:9, 12s (shot P-02).
Style: High-end coffee ad, translucent backlit morning light through a big window, alternating shallow-DOF macro and environmental medium shots, fine airborne grain. Duration: 12s. Scene: A wooden bar counter by a café window, morning light slanting in through the floor-to-ceiling glass — a cream ceramic pour-over kettle, a glass server, a white cone dripper on the counter; a blurred background of a grinder, plants, a chalkboard menu, and the occasional silhouette of a passer-by outside. Cast: Only the barista's hands appear — clean hands, sleeves rolled up. [00:00-00:04] Macro; the barista's hand pours a thin steady stream of hot water into the center of the coffee bed, the grounds blooming up into a dome, fine bubbles bursting on the surface, white steam rising in the backlit window light, the blurred window light forming a soft halo. [00:04-00:08] The camera drops to the server; deep-amber coffee drips in a continuous fine thread, swirling as it lands, the backlight passing through it into a ruby glow. In the blurred background a passer-by's silhouette drifts by outside. [00:08-00:12] Slow pull back to a medium shot revealing the full counter: the hand sets the kettle back down, steam still rising, morning light dappling the wood and the vessels. It holds on this quiet corner of the café. Audio: The thin stream of water, bursting bubbles, the crisp drip into the server, faint room tone and a distant grinder, mellow jazz piano. Constraints: Correct five-finger hand structure and a natural pouring grip; the liquid flows continuously without breaking; steam obeys hot-air rising; the background stays blurred but real and perspective-consistent; the vessel positions and materials stay consistent.
Recommended model:seedance-2
Alley Walk (POV)
UGC · POV Vlog — 9:16, 15s (shot R-01).
Style: First-person travel vlog, action-cam handheld feel, natural walking sway, wide-angle with slight distortion, auto-exposure shifting with the light, lived-in messy detail. Duration: 15s, single shot. Scene: A narrow old-town alley at dusk, weathered brick walls on both sides and balconies hung with laundry, a wet stone-paved path reflecting the warm yellow shop lights. [00:00-00:05] The camera moves forward with the footsteps, passing through a narrow alley strung with colorful cloth banners swaying overhead in the wind, light and shadow flickering fast across the lens, a person on an e-scooter brushing past on the right. [00:05-00:10] The gaze turns naturally to a lit food stall on the left, the vendor flipping food on a hot iron griddle, steam billowing up; the camera lingers briefly, oil sputtering and the food changing color. [00:10-00:15] The gaze turns back forward and keeps walking, out of the alley into an open little square, the setting sun flaring straight into the lens with glare and halation, a flock of pigeons startled up by the footsteps flying across the frame. Audio: Footsteps, market chatter, an e-scooter passing, the griddle sizzling, pigeon wingbeats, real spatial audio, no score. Constraints: Strictly first-person throughout, the filmer never appears; the sway is realistic, not exaggerated; exposure changes naturally; single shot, no cuts.
Recommended model:seedance-2
Retro GRWM
UGC · GRWM — 3:4, 12s (shot R-02).
Style: GRWM lifestyle vlog, 35mm film texture, warm-yellow morning light, slight handheld shake, obvious grain, warm faded color. Duration: 12s, 9:16 vertical. Scene: A vintage apartment bedroom — a wooden vanity, an old table lamp, curtains half open. Cast: A Latino man around 25, dark curly hair, short stubble along the jaw, a white tank top. [00:00-00:04] The camera is at the vanity; he's just woken, hair a mess, rubs his face in the mirror, yawns, reaches over and pulls the curtain open — morning light floods the whole room at once. [00:04-00:08] A few quick action cuts (about a second each): twisting open a pomade jar, working it into his hair, buttoning a shirt, spraying cologne, putting on a watch. Each is a close-up, crisp rhythm. [00:08-00:12] He steps back to the full-length mirror, fixes his collar, checks the look side-on, nods at himself in the mirror, grabs the keys off the table and walks out of frame. The camera stays on the empty room and the swaying curtain. Audio: Room tone, fabric friction, the hiss of cologne spray, keys clinking; a lazy lo-fi track. Constraints: Face and hair stay consistent across the quick cuts; each quick action is a complete beat with no dropped frames; film grain stays steady in strength.
Recommended model:seedance-2
Ask a Simple Question
UGC · Street Interview — 9:16, 10s (shot R-05).
Style: American street-interview, handheld with slight shake, natural afternoon daylight, shallow depth of field blurring the passing crowd, wide-ish medium shot, UGC feel but sharp; punchy montage editing. Duration: 10s. Scene: A busy US downtown sidewalk, afternoon, storefronts and pedestrians blurred behind. Cast: The interviewer holds a foam-covered mic, ALWAYS back to camera (only the back of the head and shoulders visible); four different young people, varied ethnicities, hair and streetwear, each facing camera in turn. Caption (persistent): "ASK A SIMPLE QUESTION" stays fixed on screen the ENTIRE clip — bold white sans-serif, centered lower third, thin black outline, correctly spelled. It never changes and is the only on-screen text. [00:00-00:03] Person 1 (young man, backwards cap, denim jacket): interviewer asks fast, "How many continents are there?" He nods confidently, leans into the mic: "...Four." Then a flicker of doubt crosses his face. [00:03-00:05] Person 2 (young woman, hoop earrings, cropped hoodie): interviewer asks, "What's seven times eight?" She freezes, eyes darting, mouths silently, then blurts "...fifty-four?" with a nervous laugh. [00:05-00:07] Person 3 (guy with glasses, flannel shirt): interviewer asks, "Name a country." Long confident pause, then dead serious: "...Africa." He nods, fully believing it. [00:07-00:10] Person 4 (girl in a beanie and oversized tee): interviewer asks, "What year did World War Two end?" She squints, thinks hard, then confidently: "...Nineteen-ninety?" Beat. She shrugs, grinning, no idea she's wrong. Cut to black. Audio: Native American-English voices, natural and casual; each wrong answer delivered with total confidence or comic hesitation; street ambience (traffic, chatter, distant music); a light rhythmic beat with a soft stinger on each cut; no other music bed. Constraints: The caption is the ONLY on-screen text, spelled exactly, unchanged the whole clip. The interviewer stays back-to-camera, face never shown. Each of the four people is clearly distinct and consistent within their own shot; natural tight lip-sync on each answer; background pedestrians never stare at the lens; keep each answer short so it lands cleanly before the cut.
Recommended model:seedance-2
The Unmasking
Cinematic · Single Shot — 16:9, 24s (shot T-01).
Style: European court masquerade, candle-gold and deep-blue night, cinematic lighting, satin and crystal texture, shallow depth of field, 4K. Duration: 24s, single shot. Scene: A palace ballroom under candle chandeliers, dozens of masked guests in finery dancing, the marble floor reflecting candlelight, deep-blue night beyond the tall windows. Cast: A white woman in an emerald satin gown, auburn hair up with a few loose curls, a gold lace mask, standing with her back to camera at the center of the floor. [00:00-00:08] The camera pushes slowly in from the ballroom entrance, threading through spinning dancers and candlelight, gradually focusing on the green-gowned figure at center with her back to us; the dancers naturally part into a line of sight. [00:08] She turns on cue — the skirt flaring into a full arc, candlelight sweeping the satin into flowing sheen, and she looks straight to camera through the gold mask. [00:08-00:20] She dances to the waltz; the camera orbits a half-circle around her, the skirt opening into a circle on each turn, the surrounding dancers blurring into points of light; she gradually dances closer to camera. [00:20] The music pauses; on cue she raises a hand and lifts the mask away — revealing her full face, a slight smile, candlelight glinting in her eyes. [00:20-00:24] Every candle in the hall brightens a notch in that instant; the camera pulls slowly back and freezes: she stands at the center of the floor holding the mask, the dancers around her still, watching. Audio: A live string waltz; the 0:08 turn and the 0:20 unmasking each land on a musical accent; skirt rustle and the crowd's faint murmur. Constraints: The 0:08 turn and 0:20 unmasking must hit precisely on the beat, within 1 second; the face is consistent before and after unmasking; the satin skirt carries real weight and inertia; the crowd never blocks the lead; no text.
Recommended model:seedance-2
Diner Flash-Mob
Music · Musical — 16:9, 10s (shot C-15).
Style: Vibrant modern movie-musical, an American retro diner, warm morning light, candy-colored art direction, smooth upbeat camerawork. Duration: 10s. Scene: A 1950s-style American diner — red booths, counter stools, a jukebox, breakfast hour. Cast: A waitress, a cook, and a dozen-plus customers of varied ethnicities; at a window booth, an old white man eating pancakes alone. [00:00-00:03] A normal morning: the waitress weaves through pouring coffee, plates clatter, chatter hums. The jukebox suddenly lights up on its own and a musical intro kicks in. [00:03-00:08] The whole diner bursts into song and dance at once: the waitress slide-steps and spins using the coffee pot as a mic, the cook drums a beat on the pans with a spatula, the customers all spin down off the stools at once into a kick-line, the camera gliding through the whole choreography in one smooth take. [00:08-00:10] The camera pushes to the window booth — the old man never looks up, focused on cutting his pancakes as if nothing is happening behind him. He notices the camera, glances up, shrugs, and keeps eating. Freeze. Audio: Upbeat jazz swing, cutlery taps locked to the beat, full-cast harmony in the dance section; in the last two seconds the music continues but at the old man's table there's only the sound of his knife and fork. Constraints: The diner set stays consistent before and after the musical burst; the group dance is synced to the beat; the old man never participates, his movements and rhythm completely unrelated to the music; the twist lands in the final 2 seconds.
Recommended model:seedance-2
Underwater Solo
Music · Underwater — 16:9, 12s (shot M-02).
Style: Artistic music video, underwater cinematography, blue-green palette, abundant bubbles and God-ray light beams, slow motion, soft and dreamlike. Duration: 12s. Scene: Deep blue underwater, sunlight piercing the surface above into radiating beams. Cast: A Black woman dancer in a long white tulle gown, deep skin tone, voluminous curly hair spreading like a cloud in the water. [00:00-00:04] Wide shot, she rises slowly from the bottom of the frame, the tulle gown and long hair unfurling in the water like slow fluid, strings of bubbles rising from her fingertips. [00:04-00:08] Medium shot, she completes a slow spin underwater, the gown forming a full circle, the light beams passing through the semi-transparent fabric into layers. The camera orbits with her spin. [00:08-00:12] She reaches up to touch the surface; the instant her fingertips break through, the surface bursts into a ring of ripples and sunlight pours down through the opening, enveloping her. It holds on her pose, head tilted up into the light. Audio: The muffled thud and bubble sounds of underwater, a low heartbeat, an ethereal female vocal and strings. Constraints: The tulle and hair move with underwater fluid resistance, not like drifting in air; bubbles rise in a consistent direction; the face stays consistent throughout.
Recommended model:seedance-2
Barrio Night
Music · Reggaeton — 16:9, 15s (shot M-04).
Style: Latin reggaeton music video, a Puerto Rican block at night, high-saturation magenta-and-cyan neon clash, humid night air, slight film grain and lens glow, commercial MV quality. Aspect: 16:9 generated, subject held mid-frame throughout. Duration: 15s. Scene: A Caribbean-style block street party at night — a sky-blue vintage convertible lowrider parked mid-street, headlights on, chrome bumper and rims catching the neon; low houses painted pink and mint on both sides, string lights along the eaves, palm silhouettes swaying above the glow; a few motorcycles parked at angles by the curb, a crowd holding up phones to the beat, the wet ground reflecting magenta neon; every sign is a solid-color neon-tube shape with no readable text. Cast: Five dancers, mostly Latino: the center is a deep-brown-skinned Latina — wet curly long hair, big gold hoop earrings, a sequined cami over low-rise cargo pants; the other four are a mixed group in oversized streetwear (bucket hats, tanks, beaded chains, skate shoes), each of different skin tone and hair. [00:00-00:05] A low wide angle glides in off the lowrider's chrome nose, the bumper reflecting flowing neon; the five dance to the dembow beat in front of the car, hips and shoulders loose but locked precisely to the beat, the crowd waving on both sides, footsteps shattering the neon reflections in the standing water. [00:05-00:10] Cut to the center dancer alone, medium: she looks back over her shoulder at camera, flips her wet hair, fingertips grazing a gold hoop, drops into a sinking perreo groove then hits a clean freeze; the motorcycle headlights flash twice behind her, the four backup dancers holding a tight groove in the blurred layer without pulling focus. [00:10-00:15] Cut to a slow orbiting wide: the five spread out around the side of the lowrider, the center sits on the hood and lifts her chin straight to camera, the other four each strike a pose beside the car; on the last beat every neon and headlight overexposes in a single flash, freezing into an album-cover composition. Audio: A reggaeton dembow beat (sinking bass, crisp drums), crowd cheers and whistles, one motorcycle-engine roar; the moves strictly synced to the beat, no dialogue. Constraints: The five dancers' skin tones, faces and outfits never swap or blend; the center's face and gold hoops stay consistent throughout; the groove stays locked to the dembow beat; the lowrider's model, color and chrome reflections stay consistent; neon is graphic light-source only, no readable text anywhere; wet-ground reflections match the light direction; formation changes are continuous, no teleporting.
Recommended model:seedance-2
Astorie canvas workflow
Every prompt in this library follows the same shot-brief structure Seedance 2.5 responds to best: a Style line that pins the look, a Duration, a Scene and Cast block, timestamped beats like [00:05-00:10] for each action, and a closing Constraints list that locks consistency — faces, wardrobe, physics, and text that must not drift.
Timestamped beats are the part worth stealing: 2.5 lands timed cues within about a second, so a 15-second prompt written as three 5-second beats comes back cut the way you scripted it. The Audio lines document the full model behavior — on Astorie, Seedance 2.5 renders are currently video-only, so treat them as pacing notes for your edit.
All clips were generated in-house with Seedance 2.5 at 720p, 16:9 unless marked 9:16, then compressed for the web. Prompts are free to reuse and remix.
Every new canvas opens with Seedance 2.5 ready to go. Paste a prompt, pick your aspect ratio and duration, and generate — free accounts include a one-time Seedance 2.5 unlock.
Related models
Related blog posts
Try this prompt on Astorie
Copy a recipe above, drop it into a node, and run it inside a full canvas workflow.