Skip to content

Vale Holdings tower on a rainy night

CEO romanceEstablishing

Night, rain, practical light, high contrast, cool tones with a warm window glow, low-angle establishing shot, wide-angle lens, vertical composition. A black glass skyscraper rises into low clouds above a wet city avenue. One office near the top is still lit gold while every floor around it is dark. Rain streaks across the frame, taxis pass below with red taillights smeared on the asphalt, and a lone figure in a tan trench coat hurries toward the revolving door. The camera tilts up slowly from the street to the lit penthouse window.

Negative: blurry, low quality, flicker, warped buildings, readable signs, logos, watermark, subtitles, daytime, sunshine

Settings: Wan 2.2 T2V-A14B, 9:16 (720×1280), 81 frames at 16 fps (about 5 s), 720p

Wan 2.2 was trained with labels for light, framing and tone. So the comma-separated tags up front really steer it. One slow tilt up keeps the A14B model from adding extra motion.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

"You're late, Miss Brooks"

CEO romanceClose-up

Night, practical light from a desk lamp, side light, soft light, low saturation, cool tones, close-up, eye-level, center composition, long-focus lens. A 32-year-old CEO with a sharp jaw, slicked-back black hair, cold dark eyes, a charcoal three-piece suit and silver cufflinks sits behind a glass desk in a penthouse office, rain running down the window behind him. He doesn't look up from the document in his hand. After a beat his eyes lift to the camera, unimpressed. [CEO, low calm voice, slow, American English]: "You're late, Miss Brooks." Sound: rain against glass, a clock ticking, faint traffic far below. No background music. Generate single shot.

Negative: smiling, cartoon, distorted face, extra fingers, on-screen text, subtitles, watermark, blurry

Settings: Wan 2.7 text-to-video, 9:16, 5 s, 1080p, audio generated

Alibaba's voice formula (line + emotion + tone + speed + accent) gives steadier delivery than a bare quote. 'Generate single shot.' stops Wan 2.7 from cutting away mid-line.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

The resignation letter: the full scene

CEO romanceOver-the-shoulder

Night, practical light, soft side light, warm tones on her face and cool tones on his, over-the-shoulder shot from behind the CEO, medium close-up on her, medium-focus. In the foreground, the shoulder and slicked-back black hair of a man in a charcoal three-piece suit, out of focus. Facing him across a glass desk, a 24-year-old executive assistant with chestnut hair in a low ponytail, a cream silk blouse and a thrifted tan trench coat. She steps forward and slaps a white envelope onto the desk, chin trembling but eyes steady. [Assistant, shaky but defiant, quick]: "I quit. Find someone else to humiliate." She holds his gaze. Immediately, the man slowly stands, his shadow falling across her, and she has to tilt her head up to keep looking at him. Fixed camera. Sound: the slap of paper on glass, rain on the window, her uneven breathing. A low, tense cello note rises at the very end. Generate single shot.

Negative: camera shake, outfit changes, extra people, distorted hands, writing on the envelope, subtitles, watermark, blurry

Settings: Wan 2.7 image-to-video with first and last frame, 9:16, 10 s, 1080p, audio generated. First frame: her at the desk, envelope in hand. Last frame: him standing over her.

Giving Wan 2.7 both the first and last frame pins down who stands where. The 10 seconds then go to acting, not wandering. 'Immediately,' is Alibaba's linking word, so her line and his reaction stay separate.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Silver cufflink, signed contract

CEO romanceInsert

The hand signs the last line of the contract with the fountain pen in one smooth stroke, then lifts the pen away. The silver cufflink catches a brief glint of lamp light. Slow, deliberate motion. Fixed camera.

Negative: readable text, letters, distorted fingers, six fingers, blurry, flicker, watermark

Settings: Wan 2.2 I2V-A14B, 9:16 (720×1280), 81 frames at 16 fps (about 5 s), 720p. Start image: top-down close-up of a man's hand in a charcoal sleeve with a silver cufflink, a fountain pen over the last page of a contract, warm desk-lamp light.

In image-to-video, Alibaba says the image sets the subject, scene and style. So the prompt only covers motion and camera. Inserts like this are the cheapest way to hide a cut.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Chase down the executive floor (multi-shot)

CEO romanceTracking

A tense office-romance beat in four shots: the assistant storms out and the CEO follows her down the executive floor. Image 1 is the CEO: a 32-year-old man with a sharp jaw, slicked-back black hair, cold dark eyes, a charcoal three-piece suit and silver cufflinks. Image 2 is the assistant: a 24-year-old woman with chestnut hair in a low ponytail, a cream silk blouse and a thrifted tan trench coat. Night, glass corridors, cool fluorescent light mixed with warm desk lamps, vertical framing. Shot 1 [0–4 s] Tracking shot from behind: the camera follows Image 2 as she walks fast down a long glass corridor, clutching her bag strap, staff turning to watch. Shot 2 [4–7 s] Hard cut. Medium shot: Image 1 steps out of his office and strides after her, buttoning his jacket, the camera tracking backward in front of him. Shot 3 [7–11 s] Hard cut. Over-the-shoulder from behind Image 2 at the elevator. Image 1 catches the closing doors with one hand. [CEO, quiet, controlled]: "We're not finished." Shot 4 [11–15 s] Hard cut. Close-up on Image 2 inside the elevator, breath held, eyes darting to his hand on the door. [Assistant, whispering]: "We never started." Sound: hurried heels on marble, a muffled phone ringing, the elevator chime. A low pulsing synth score that drops out on the last line.

Settings: Wan 3.0 reference-to-video, 9:16, 15 s, 1080p, audio on. References: Image 1 = CEO character sheet, Image 2 = assistant character sheet.

Wan 3.0 reads 'Image 1' and 'Image 2' as your uploads, in order. So you can name each character by role without describing the face again. Timestamps plus 'Hard cut' give you real cuts in one 15-second clip.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

The ex-wife walks into the gala

RevengeWide

Night, chandeliers, warm practical light, high contrast, wide shot, low angle, symmetrical composition, vertical frame. A packed gala ballroom with gold walls and a string quartet. The double doors at the far end swing open and a 28-year-old woman with a sleek platinum bob, red lipstick and a black satin dress walks in alone. Conversations stop table by table as she passes, heads turning in a wave. She walks straight toward the camera, unhurried, while the crowd parts. The camera holds still and lets her grow in the frame. Sound: the quartet falters and stops mid-phrase, glasses clink, a murmur spreads, her heels click on marble. Generate single shot.

Negative: running, waving, extra limbs, melted faces in the crowd, readable signs, watermark, blurry

Settings: Wan 2.7 text-to-video, 9:16, 10 s, 1080p, audio generated

A fixed camera with her walking toward the lens still works on a phone. She fills more of the frame every second. When the quartet stops, Wan's audio becomes part of the story.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

"Hello, Grant. Miss me?"

RevengeClose-up

She tilts her head slightly and a slow, cold smile spreads across her red lips. Her eyes stay fixed on someone just off camera. [Woman, velvet-soft, amused, slow]: "Hello, Grant. Miss me?" The camera pushes in slowly toward her face. Sound: muffled party chatter behind her, a glass set down hard. No background music.

Negative: exaggerated grin, distorted lips, rapid blinking, subtitles, text, watermark

Settings: Wan 2.6 image-to-video, 9:16, 5 s, 1080p, audio generated, multi-shot off. Start image: close-up of the woman with the platinum bob, red lipstick and black satin dress at the gala.

In image-to-video the still sets her look. So spend the words on the smile, the eyeline and one slow push-in. One short line per five-second clip keeps Wan's lip sync on track.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Diamond ring drops into champagne

RevengeInsert

Warm ballroom light, shallow depth of field, high contrast, insert shot, macro close-up, side angle. A woman's hand with pink tweed cuffs slides an oversized diamond ring off her finger and lets it drop into a flute of champagne. The ring sinks through the bubbles in slow motion and settles at the bottom of the glass. Chandelier bokeh in the background. Fixed camera.

Negative: blurry, deformed fingers, extra fingers, liquid splashing everywhere, text, watermark, flicker

Settings: Wan 2.2 T2V-A14B, 9:16 (720×1280), 81 frames at 16 fps (about 5 s), 720p. Optional: 1 interpolated frame for smoother slow motion.

Wan handles one clear physical action well: a single drop and a slow sink. Wan 2.2 runs at 16 fps by default. Frame interpolation (FILM or RIFE on fal) smooths the slow motion.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Grant's smile dies on a text message

RevengeMedium

Night, mixed light from chandeliers and a phone screen, soft light, warm tones, medium shot, eye-level, right-heavy composition. A 35-year-old man in a navy suit with a smug half-smile stands at a cocktail table holding a glass of whiskey. His phone buzzes; he glances at it, still smirking. Then he reads it again. The smile drains from his face, his jaw tightens, and he sets the glass down without looking. Sound: the phone buzz, the clink of the glass, party noise fading into a low hum, one low piano note. No dialogue. Generate single shot.

Negative: cartoon expression, readable phone screen, extra fingers, warped face, watermark, blurry

Settings: Wan 2.7 text-to-video, 9:16, 5 s, 720p, audio generated

'No dialogue.' is Alibaba's switch to stop Wan from inventing lines in a silence. The turn is written as visible beats (glance, reread, jaw). Wan follows actions better than an adjective like 'shocked.'

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Courthouse steps showdown (multi-shot)

RevengeOver-the-shoulder

A divorce-court revenge beat in three shots. Overcast daylight, low saturation, cool tones, vertical framing. A 28-year-old woman with a sleek platinum bob, red lipstick and a black satin coat faces her ex-husband, a 35-year-old man in a navy suit with a smug half-smile, on the stone steps of a courthouse while reporters crowd below. First shot [0-5s] Over-the-shoulder from behind the man, his navy shoulder in the foreground, the woman two steps above him looking down. [Ex-husband, condescending]: "You walked away with nothing, Claire." Second shot [5-10s] Hard cut, reverse over-the-shoulder from behind her platinum hair onto his face. She holds up a thin folder. [Woman, quiet, final]: "I walked away with your company." Third shot [10-15s] Hard cut, medium close-up on the man as camera flashes go off and his smile cracks. Sound: shutter clicks, reporters shouting, wind. Tense strings build under the last shot.

Negative: different faces between shots, extra people on the steps, warped hands, readable text, watermark, subtitles

Settings: Wan 2.6 text-to-video, 9:16, 15 s, 1080p, audio generated, multi-shot on, prompt expansion on

This uses Wan 2.6's multi-shot format: overall description first, then timed shots. It stays under the 1,500-character prompt limit. On fal, multi-shot only works with prompt expansion on.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Silvercrest forest under a full moon

WerewolfEstablishing

Night, moonlight, cool tones, low saturation, rim light on the trees, high-angle establishing shot, aerial view, wide-angle lens, vertical composition. A dense pine forest wrapped in drifting mist under a huge full moon. In a clearing sits a log pack lodge with warm firelight in its windows. Mist rolls slowly between the trunks. The camera descends slowly toward the lodge roof. Far off, a pair of amber eyes flickers between the trees and vanishes.

Negative: daylight, blurry, flicker, distorted trees, cartoon, text, watermark

Settings: Wan 2.2 T2V-A14B, 9:16 (720×1280), 81 frames at 16 fps (about 5 s), 720p

Ending on a tiny reveal (eyes in the trees) turns an establishing shot into a hook. Wan 2.2 holds one slow descent well. Don't add a second camera move in five seconds.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

The Alpha's eyes turn amber

WerewolfClose-up

His eyes narrow and the irises shift from dark brown to glowing amber, the glow spreading outward from the pupils. The scar through his left eyebrow tightens as he frowns. A slow breath fogs in the cold air. Moonlight flickers across his face as branches sway. The camera pushes in very slowly.

Negative: whole face glowing, deformed eyes, cross-eyed, extra eyes, blurry, flicker, watermark

Settings: Wan 2.2 I2V-A14B, 9:16 (720×1280), 81 frames at 16 fps (about 5 s), 720p. Start image: moonlit close-up of a 30-year-old man with shoulder-length dark hair, a scar through his left eyebrow and a black leather jacket.

The change is gradual and has a direction ('spreading outward from the pupils'). That gives Wan something to animate across 81 frames instead of a sudden pop.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Ivy runs from the pack

WerewolfTracking

Night, moonlight through mist, cool tones, backlight, tracking shot, follow-cam, eye-level, vertical framing. A 22-year-old woman with freckles, auburn curls and an oversized red flannel shirt runs barefoot through a misty pine forest, glancing back over her shoulder. The camera tracks alongside her at running speed, branches whipping past the lens. Behind her, low shapes move between the trees. Sound: her ragged breathing, snapping twigs, a long wolf howl answered by others, drums like a heartbeat. No dialogue. Generate single shot.

Negative: slow motion, wolves in clear view, extra limbs, warped legs, blurry, text, watermark

Settings: Wan 2.7 text-to-video, 9:16, 8 s, 720p, audio generated

Red flannel against blue moonlight gives Wan a high-contrast target. She stays easy to see at 720p. Calling the wolves 'low shapes' avoids the anatomy mistakes full creatures cause in fast motion.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

"I reject you, Ivy Hart"

WerewolfMedium

He steps toward her, then stops an arm's length away. His face is hard but his hands are clenched at his sides. [Alpha, deep, rough, forcing the words]: "I, Kade Thorne, reject you, Ivy Hart." She flinches as if struck and takes one step back. Fixed camera. Sound: wind through the pines, a crackling bonfire behind them, a murmur from the unseen pack. No background music.

Negative: smiling, hugging, extra people in front, distorted faces, subtitles, watermark

Settings: Wan 2.7 image-to-video, 9:16, 6 s, 1080p, audio generated. Start image: medium two-shot of the Alpha (leather jacket, scarred eyebrow) and Ivy (auburn curls, red flannel) by a bonfire.

The rejection line is the genre's big beat, so it gets its own clip. Alibaba's guide says to tie speech to a visible action. The clenched hands and her flinch give the line something to land on.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Rejection ceremony to hidden-power reveal (multi-shot)

WerewolfWide

A werewolf-romance turning point in five shots. Night, full moon, bonfire light, cool blue against warm orange, vertical framing. Image 1 is the Alpha: a 30-year-old man with shoulder-length dark hair, a scar through his left eyebrow and a black leather jacket. Image 2 is Ivy: a 22-year-old woman with freckles, auburn curls and an oversized red flannel shirt. Image 3 is the forest clearing with the pack lodge. Shot 1 [0–4 s] Wide shot of the clearing from Image 3: forty pack members in a circle around a bonfire, Image 1 at the center, Image 2 pushed forward alone. Shot 2 [4–8 s] Hard cut. Medium shot of Image 1. [Alpha, cold]: "I reject you." Shot 3 [8–12 s] Hard cut. Close-up of Image 2, tears on her freckles, her eyes suddenly glowing silver. Shot 4 [12–16 s] Hard cut. Wide shot: a ring of silver light bursts out from Image 2 and every pack member drops to one knee, except Image 1. Shot 5 [16–20 s] Hard cut. Close-up of Image 1, amber eyes wide, his hand shaking. Sound: bonfire crackle, gasps, a deep rumbling hum when the light bursts. A choir swells on shot 4, then silence on shot 5.

Settings: Wan 3.0 reference-to-video, 9:16, 20 s, 1080p, audio on. References: Image 1 = Alpha character sheet, Image 2 = Ivy character sheet, Image 3 = clearing location still.

Wan 3.0 takes a location as a reference image too, so the clearing looks the same every episode. The silent close-up at the end is your cliffhanger frame.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Carriage arrives at Ravensworth in the rain

HistoricalEstablishing

Overcast dusk, rain, low saturation, cool tones with warm candlelit windows, establishing shot, wide-angle lens, eye-level, vertical composition. A grand Regency-era stone manor at the end of a gravel drive lined with bare trees. A black horse-drawn carriage rolls toward the front steps, wheels cutting through puddles, lanterns swinging on its sides. A footman with an umbrella hurries down the steps. Rain falls steadily. The camera pushes in slowly toward the carriage door.

Negative: modern cars, power lines, street lamps, blurry, deformed horses, extra legs, text, watermark, flicker

Settings: Wan 2.2 T2V-A14B, 9:16 (720×1280), 81 frames at 16 fps (about 5 s), 720p

The negative prompt removes modern details Wan sometimes adds to period scenes. A carriage at a walk hides horse-leg glitches better than a gallop.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

The duke asks for a dance

HistoricalOver-the-shoulder

Candlelight, warm tones, soft light, rim light, over-the-shoulder shot from behind the lady, medium close-up on him, long-focus lens. In the foreground, the pinned-up dark curls and pearl drop earring of a 20-year-old Regency lady in a pale blue empire-waist gown. Facing her in a crowded candlelit ballroom, a 29-year-old duke in a black tailcoat and white cravat with cold gray eyes bows slightly and offers a gloved hand. [Duke, low, formal, a little bored]: "Lady Eleanor. I believe this dance is mine." Couples whirl past behind him. Sound: a string ensemble playing a waltz, rustling silk, murmured gossip. Generate single shot.

Negative: modern clothing, warped hands, extra fingers, melting faces in the background, subtitles, watermark

Settings: Wan 2.7 text-to-video, 9:16, 8 s, 1080p, audio generated

Name what's in the foreground (her curls and earring) to get a true over-the-shoulder shot. Otherwise Wan gives you a flat two-shot. The long-focus lens squeezes the ballroom so the vertical frame stays clean.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

A sealed letter changes hands

HistoricalInsert

Candlelight, warm tones, shallow depth of field, insert shot, close-up, top-down angle. A woman's white-gloved hand slides a folded letter sealed with dark red wax across a polished wooden table toward a man's black-gloved hand. His fingers stop on the seal. The candle flame flickers. Fixed camera.

Negative: readable handwriting, modern objects, deformed fingers, blurry, watermark

Settings: Wan 2.2 TI2V-5B, 9:16 (704×1280), 121 frames at 24 fps (about 5 s), 720p

The 5B model fits on one consumer GPU, and it does best on simple inserts like this. Its vertical 720p size is 704×1280, not 720×1280.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Reading his letter by the window

HistoricalMedium

She breaks the wax seal and unfolds the letter, reading quickly. Her lips part. Halfway down the page her hand begins to tremble, and a single tear runs down her cheek past her pearl earring. She presses the letter to her chest and looks out at the rain-streaked window. The camera pushes in slowly from a medium shot to a medium close-up. Sound: rain on glass, a crackling fireplace, paper rustling. A soft, melancholy solo piano.

Negative: readable writing on the letter, loud sobbing, distorted face, extra fingers, watermark

Settings: Wan 2.6 image-to-video, 9:16, 10 s, 1080p, audio generated, multi-shot off. Start image: the Regency lady (dark curls pinned up, pale blue gown, pearl earrings) at a candlelit window holding a sealed letter.

Ten seconds is enough for an emotional build if you write it in order: break, read, tremble, tear, look away. Leaving multi-shot off keeps Wan 2.6 on one continuous take.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

The first waltz (two reference images)

HistoricalTracking

Image 1 is a 29-year-old duke in a black tailcoat and white cravat with cold gray eyes. Image 2 is a 20-year-old lady with dark curls pinned up, a pale blue empire-waist gown and pearl drop earrings. A candlelit Regency ballroom, warm tones, soft rim light, vertical framing. Image 1 and Image 2 waltz in the center of the floor while other couples blur past. The camera circles them slowly in one continuous tracking shot. Image 2 looks up at Image 1, defiant. [Image 2, quiet, sharp]: "You don't even like dancing." Immediately, [Image 1, almost smiling]: "I don't. I like winning." He spins her once and pulls her close. Sound: strings playing a waltz, silk swishing, footsteps on parquet. Generate single shot.

Negative: dress changing color, modern clothes, tangled limbs, extra arms, warped faces, subtitles, watermark

Settings: Wan 2.7 reference-to-video, 9:16, 10 s, 1080p, audio generated. References: Image 1 = duke character sheet, Image 2 = lady character sheet.

Wan 2.7 reference-to-video stops at 10 seconds. So the scene is one circling take with two lines. Label each speaker by image number and link the lines with 'Immediately,' per Alibaba's multi-character dialogue rules.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Lighthouse town in the storm

ThrillerEstablishing

Night, storm, hard light from a rotating lighthouse beam, high contrast, cool tones, low saturation, establishing shot, wide-angle lens, high angle, vertical composition. A small coastal town of wooden houses clings to black cliffs. Waves smash against the rocks and spray drifts across the frame. The lighthouse beam sweeps across wet rooftops and an empty road where a police car sits with its lights off. The camera holds still as the beam passes twice.

Negative: daylight, sunshine, blurry, flicker, distorted buildings, text, watermark

Settings: Wan 2.2 T2V-A14B, 9:16 (720×1280), 81 frames at 16 fps (about 5 s), 720p

A moving light over a still camera gives you motion without the camera drifting. Wan 2.2 renders that cleanly at 16 fps.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Unknown caller at 3 a.m.

ThrillerClose-up

Night, practical light from a phone screen, cool blue tones, low key, close-up, eye-level, center composition. A 34-year-old detective with short black hair, tired eyes and an olive raincoat sits in a dark parked car, rain on the windshield behind her. Her phone lights up her face; she answers and listens. Her eyes widen slightly. [Distorted male voice through the phone, slow, calm]: "You searched the wrong lighthouse, Detective." She lowers the phone slowly. Sound: rain drumming on the car roof, a low electronic drone. Generate single shot.

Negative: readable phone screen, smiling, distorted face, extra fingers, subtitles, watermark, blurry

Settings: Wan 2.7 text-to-video, 9:16, 6 s, 1080p, audio generated

Calling the voice 'distorted' and 'through the phone' stops Wan from putting the line in her mouth. A phone screen as the only light is a reliable thriller look on Wan.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Interrogation room with two reference images

ThrillerOver-the-shoulder

Image 1 is a 34-year-old detective with short black hair, tired eyes and an olive raincoat. Image 2 is a soft-spoken 31-year-old man with wire-rimmed glasses and a gray cardigan. A cramped interrogation room, one hanging lamp, hard top light, high contrast, cool tones, vertical framing. Over-the-shoulder shot from behind Image 1, her shoulder and short black hair in the foreground, Image 2 seated across the metal table. She slides a photo across the table. Image 2 looks down at it, then back up, perfectly calm. [Image 2, gentle, almost apologetic]: "I've never seen her before." Immediately, [Image 1, flat, tired]: "Then why did you take her picture?" Image 2 blinks once and adjusts his glasses. Fixed camera. Sound: a buzzing lamp, a chair creak, distant thunder. No background music. Generate single shot.

Negative: readable photo, extra people, warped hands, glasses changing shape, subtitles, watermark

Settings: Wan 2.7 reference-to-video, 9:16, 10 s, 1080p, audio generated. References: Image 1 = detective character sheet, Image 2 = suspect character sheet.

A fixed over-the-shoulder is a safe dialogue setup on Wan 2.7 reference-to-video. One face is mostly hidden, so the model only has to hold one identity. The small glasses adjustment sells the cliffhanger.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Motel key on a wet floor

ThrillerInsert

Water drips steadily onto the brass key and its plastic number tag, rippling the puddle around it. A flashlight beam slides across the floor and stops on the key. A gloved hand enters the frame and picks it up. Fixed camera, slow motion.

Negative: readable numbers, distorted hand, extra fingers, blurry, flicker, watermark

Settings: Wan 2.2 I2V-A14B, 9:16 (720×1280), 81 frames at 16 fps (about 5 s), 720p. Start image: a brass motel key with a plastic tag lying in a puddle on cracked tile, faint flashlight glow.

Design the still first in an image model, since it handles the key tag better. Then let Wan animate only the drip, the beam and the hand. That's Alibaba's split for image-to-video: image for look, prompt for motion.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Following the suspect to the lighthouse (30-second sequence)

ThrillerTracking

A 30-second thriller sequence in six shots. Night, storm, the lighthouse beam as the main light source, cool tones, vertical framing. Image 1 is the detective: a 34-year-old woman with short black hair, tired eyes and an olive raincoat. Image 2 is the suspect: a soft-spoken 31-year-old man with wire-rimmed glasses and a gray cardigan. Image 3 is the lighthouse and cliff path. Shot 1 [0–5 s] Wide shot of the cliff path from Image 3, Image 2 walking toward the lighthouse with a lantern. Shot 2 [5–10 s] Hard cut. Tracking shot behind Image 1 as she follows at a distance, keeping low behind rocks. Shot 3 [10–14 s] Hard cut. Close-up of Image 1 wiping rain from her eyes, flashlight held low and switched off. Shot 4 [14–20 s] Hard cut. Tracking shot following Image 2 up the spiral stairs inside the lighthouse, lantern swinging. Shot 5 [20–25 s] Hard cut. Medium shot at the top: Image 2 stops, turns toward the stairwell and smiles. [Suspect, calm]: "You can come up, Nora." Shot 6 [25–30 s] Hard cut. Close-up of Image 1 frozen on the stairs, the lighthouse beam sweeping across her face. Sound: howling wind, waves, footsteps on iron stairs, a creaking lantern. A low ticking pulse that stops dead after his line.

Settings: Wan 3.0 reference-to-video, 9:16, 30 s, 720p, audio on. References: Image 1 = detective character sheet, Image 2 = suspect character sheet, Image 3 = lighthouse location still.

Thirty seconds is Wan 3.0's limit, so this is a whole episode beat in one pass. Draft at 720p ($0.10 per second at fal's list price). Then re-run only the keeper at 1080p.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

The mustard bottle betrayal

ComedyMedium

Daylight, bright warm tones, soft light, medium shot, eye-level, center composition. A 27-year-old food-truck owner with curly hair and a mustard-stained apron leans out of his truck window, grinning at a customer off camera, and squeezes a giant mustard bottle over a hot dog with a flourish. The bottle sputters and blasts mustard straight onto his own face. He freezes, blinks, keeps grinning. [Food-truck owner, cheerful, fast]: "Signature move. Totally on purpose." Sound: a loud squelch, a customer snorting with laughter, a sizzling grill. Light ukulele music. Generate single shot.

Negative: distorted face, extra hands, readable truck signage, watermark, blurry

Settings: Wan 2.7 text-to-video, 9:16, 6 s, 720p, audio generated

Comedy on Wan works best as one clear physical gag plus a reaction. Asking him to 'keep grinning' after the blast gives the joke a beat of stillness before the line.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Wedding venue meltdown in one wide

ComedyWide

Daylight, bright, high saturation, warm tones, wide shot, high angle, vertical composition. A garden wedding venue twenty minutes before the ceremony: a toppled flower arch, a ring bearer chasing a runaway goose across the lawn, two caterers carrying a five-tier cake in opposite directions. In the middle of it all, a 26-year-old wedding planner with a headset and a color-coded clipboard turns in a slow circle, taking it in. The camera holds still from above. [Wedding planner, into her headset, very calm]: "Code red. Everything is code red." Sound: a honking goose, kids shrieking, a string quartet tuning badly.

Negative: melted crowd faces, extra limbs, warped cake, text, watermark

Settings: Wan 2.6 text-to-video, 9:16, 10 s, 1080p, audio generated, multi-shot off

A high-angle wide lets several small gags play at once. The vertical frame stacks them top to bottom. Keep each background action to a few words so Wan doesn't merge them.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

The planner's eye twitch

ComedyClose-up

She stares straight ahead without blinking. Her left eye starts to twitch, once, twice, then faster. Her grip tightens on the clipboard until her knuckles whiten. She inhales slowly through her nose and forces a huge, fake smile. The camera pushes in slightly. Subtle, controlled motion.

Negative: both eyes twitching, distorted face, manic expression, blurry, flicker, watermark

Settings: Wan 2.2 I2V-A14B, 9:16 (720×1280), 81 frames at 16 fps (about 5 s), 720p. Start image: close-up of the wedding planner with her headset and color-coded clipboard, bright garden behind her.

Wan animates small face movements well when you count them out ('once, twice, then faster'). The negative prompt keeps the twitch to one eye, which is funnier and less strange.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

A taco lands on the wedding cake

ComedyInsert

Daylight, bright warm tones, insert shot, close-up, eye-level, shallow depth of field. A pristine white five-tier wedding cake on a stand. A taco tumbles into frame from above, spinning, and lands right on top, scattering lettuce and salsa over the white frosting. The tiny bride-and-groom topper tips over. Fixed camera.

Negative: blurry, collapsing cake, text, watermark, human faces

Settings: Wan 2.2 TI2V-5B, 9:16 (704×1280), 121 frames at 24 fps (about 5 s), 720p

Prop gags without faces suit the smaller 5B model. Its 24 fps output keeps the tumble smooth.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Food truck vs. wedding: the goose chase (multi-shot)

ComedyTracking

A slapstick rom-com beat in four shots. Bright afternoon, high saturation, warm tones, vertical framing. A 27-year-old food-truck owner with curly hair and a mustard-stained apron and a 26-year-old wedding planner with a headset and a color-coded clipboard chase a runaway goose that has stolen the ring box. First shot [0-4s] Tracking shot: the camera runs alongside the planner as she sprints across the lawn in heels, clipboard flapping. [Planner, out of breath]: "Get. The. Goose." Second shot [4-8s] Hard cut, tracking shot from behind the food-truck owner as he vaults a hay bale and misses the goose completely. Third shot [8-12s] Hard cut, low-angle medium shot: they dive at the goose from opposite sides and crash into each other. Fourth shot [12-15s] Hard cut, close-up of the two of them lying in the grass face to face, the ring box between them, both suddenly quiet. Sound: goose honks, thuds, guests gasping. Fast pizzicato strings that cut out for the final close-up.

Negative: injury, extra legs, warped goose, outfits changing between shots, text, watermark

Settings: Wan 2.6 text-to-video, 9:16, 15 s, 1080p, audio generated, multi-shot on, prompt expansion on

The rom-com trick is cutting the music for the face-to-face moment. Wan's audio follows written cues like 'cut out for the final close-up.' Repeating outfit details in the overall description keeps both leads matching across four cuts.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

The floating citadel at dawn

FantasyEstablishing

Dawn, warm golden light, rim light, soft haze, high saturation, establishing shot, extreme wide shot, low angle, vertical composition. A white stone citadel with slender spires floats above a sea of clouds, waterfalls spilling off its edges into mist. Chains of glowing blue runes orbit slowly around the lowest tower. A flock of birds crosses in front of the sun. The camera tilts up slowly from the clouds to the highest spire.

Negative: blurry, flicker, warped architecture, modern buildings, text, watermark

Settings: Wan 2.2 T2V-A14B, 9:16 (720×1280), 81 frames at 16 fps (about 5 s), 720p

Tall fantasy buildings suit a 9:16 frame, and a tilt-up uses all that height. The waterfalls give Wan constant motion to render.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

The rune ignites on her wrist

FantasyInsert

The rune tattoo on her right wrist begins to glow, faint at first, then bright blue, light running along the lines like ink catching fire. Her fingers curl. Tiny sparks drift up from her skin. The camera stays fixed. Sound: a low rising hum, a crackle like static, her sharp intake of breath. No dialogue. Generate single shot.

Negative: whole arm glowing, extra fingers, burnt skin, blurry, text, watermark

Settings: Wan 2.7 image-to-video, 9:16, 5 s, 1080p, audio generated. Start image: close-up of a pale wrist with a blue rune tattoo and the frayed edge of a moss-green cloak.

Wan 2.7 adds synced sound to image-to-video, so a magic insert gets its hum and crackle in the same pass. Keep the effect on the tattoo, or the glow spreads over the whole frame.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

The prince kneels with his sword

FantasyMedium

Dusk, cool tones with warm rim light, soft light, medium shot, low angle, center composition. In a forest of glowing translucent crystal trees, a 25-year-old prince in dark engraved armor with a thin gold circlet and storm-gray eyes drops to one knee and lays his sword flat on the ground before him. He looks up at someone just above the camera. [Prince, hoarse, sincere]: "My kingdom hunted you. I won't." Sound: crystal leaves chiming in the wind, armor creaking. A quiet choir enters. Generate single shot.

Negative: smiling, cartoon, deformed armor, extra hands, warped sword, subtitles, watermark

Settings: Wan 2.7 text-to-video, 9:16, 8 s, 1080p, audio generated

A low angle with his eyeline just above the lens puts the viewer in the heroine's place. That's a classic vertical-drama move. Short, sincere lines land better than speeches with Wan's audio.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Sera faces the shadow beast

FantasyWide

Night, cool blue tones, backlight, volumetric light through crystal trees, wide shot, eye-level, vertical composition. A 19-year-old woman with a silver-white braid, a glowing blue rune tattoo on her right wrist and a worn moss-green cloak stands alone in a crystal-forest clearing. Across from her, a huge shadow beast made of smoke with glowing red eyes rises from the ground. She raises her right hand and blue light flares from the rune, pushing the smoke back. The camera holds still. Sound: a deep growl, rushing wind, a bright magical crack. Epic drums build.

Negative: detailed monster anatomy, gore, extra limbs, blurry, flicker, text, watermark

Settings: Wan 2.6 text-to-video, 9:16, 10 s, 1080p, audio generated, multi-shot off

Making the beast smoke and eyes avoids the anatomy errors big creatures cause, and it's scarier anyway. One push of light is one action, which Wan 2.6 can carry for ten seconds.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Blade at his throat: first meeting (video + image refs)

FantasyOver-the-shoulder

A fantasy-romance first meeting in three shots. Dusk in a crystal forest, cool blue light with warm rim light, vertical framing. Video 1 is a turntable reference of Sera: a 19-year-old woman with a silver-white braid, a glowing blue rune tattoo on her right wrist and a worn moss-green cloak. Image 1 is the prince: a 25-year-old man in dark engraved armor with a thin gold circlet and storm-gray eyes. Shot 1 [0–5 s] Over-the-shoulder from behind Image 1, his armored shoulder in the foreground, as the woman from Video 1 steps out from behind a crystal tree holding a short blade. Shot 2 [5–10 s] Hard cut. Reverse over-the-shoulder from behind her silver-white braid: she presses the blade to his throat. [Sera, low, steady]: "Your father burned my village." Shot 3 [10–15 s] Hard cut. Close-up of Image 1, not resisting. [Prince, quiet]: "Then you'll want to hear what he's planning next." Sound: chiming crystal leaves, a blade sliding free, her breathing. No background music until the last two seconds, then one low string note.

Settings: Wan 3.0 reference-to-video, 9:16, 15 s, 1080p, audio on. References: Video 1 = a 5 s turntable of Sera (a slow 180° orbit on a gray backdrop, made earlier with image-to-video), Image 1 = prince character sheet. Input video seconds are billed on top of output.

A short turntable clip (the character turning in place) shows Wan 3.0 Sera's face, braid and cloak from every side. That matters when shot 2 sees her from behind. Videos and images are numbered separately, so 'Video 1' and 'Image 1' don't clash.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

The Miller porch at dusk

Family dramaEstablishing

Dusk, warm porch light, soft light, low saturation, establishing shot, eye-level, vertical composition. A small white clapboard house on a quiet street, peeling paint on the porch rail, one warm light on over the door. A bicycle lies on the lawn. Wind moves the porch swing gently. The camera pushes in slowly toward the front door.

Negative: blurry, flicker, distorted house, text, watermark, people

Settings: Wan 2.2 TI2V-5B, 9:16 (704×1280), 121 frames at 24 fps (about 5 s), 720p

Family dramas live in small, lived-in places. The 5B model handles calm exteriors well. The bike on the lawn says a teenager lives here, without a word.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Bills on the kitchen table

Family dramaMedium

Night, practical light from a single pendant lamp, soft light, warm tones, low saturation, medium shot, eye-level, left-heavy composition. A 45-year-old ER nurse with a graying bun and navy scrubs sits at a cluttered kitchen table, rubbing her temples over a stack of overdue bills. Behind her, a 16-year-old girl in an oversized hoodie with braces and headphones around her neck appears in the doorway and stops. [Teen girl, quiet, guilty]: "Mom... I need to tell you something." The mother slowly turns around. Sound: refrigerator hum, a dripping faucet, the scrape of a chair. No background music. Generate single shot.

Negative: readable bills, distorted faces, extra fingers, cheerful bright lighting, subtitles, watermark

Settings: Wan 2.7 text-to-video, 9:16, 8 s, 1080p, audio generated

Put the second character in the doorway behind the first. That gives depth without an awkward side-by-side. Ending on the mother turning around gives you a built-in cut point.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Grandpa at the door: "Hello, Gracie"

Family dramaClose-up

He lifts his eyes from the doormat to her face and tries to smile; it doesn't quite hold. His hand tightens on the cane. [Old man, gravelly, unsure]: "Hello, Gracie." Rain drips from the porch roof behind him. The camera pushes in slowly. Sound: rain, a screen door creaking, a dog barking far away. No background music.

Negative: broad smile, distorted face, extra fingers, subtitles, watermark, blurry

Settings: Wan 2.7 image-to-video, 9:16, 5 s, 1080p, audio generated. Start image: close-up of a 70-year-old man in a brown cardigan with a walking cane at a rainy front door.

Two words of dialogue, one failed smile and a push-in are about all five seconds can hold. The failed smile does more than a longer line would.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Alone at the end of the driveway

Family dramaWide

Rain falls steadily. The old man stands still at the end of the driveway, looking at the lit window of the house. After a moment he turns and walks away slowly, leaning on his cane. The porch light flickers. The camera holds still.

Negative: running, blurry, flicker, extra legs, distorted figure, watermark

Settings: Wan 2.2 I2V-A14B, 9:16 (720×1280), 81 frames at 16 fps (about 5 s), 720p. Start image: wide night shot of a small house with one lit window, a lone figure in a brown cardigan with a cane at the end of the driveway in the rain.

A small figure in a tall vertical frame reads as loneliness. A slow walk is easier for Wan to keep steady at 16 fps than a faster one.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.

Hospital corridor: the call nobody wants (first and last frame)

Family dramaTracking

One continuous tracking shot, no cuts. Fluorescent hospital light, cool tones, low saturation, vertical framing. The camera tracks backward in front of a 45-year-old ER nurse with a graying bun and navy scrubs as she walks fast down a hospital corridor reading a chart, colleagues nodding as she passes. A nurse at the station calls out to her. [Station nurse, urgent]: "Grace, your father's in bay four." She stops dead. The chart slips in her hands. She turns and walks, then runs, toward bay four, the camera now following behind her. She pulls the curtain open and freezes. Sound: monitors beeping, a PA announcement, sneakers squeaking on linoleum, her breathing getting faster. A low pulse like a heartbeat.

Settings: Wan 3.0 image-to-video with first and last frame, 9:16, 20 s, 1080p, audio on. First frame: the nurse walking with a chart. Last frame: the nurse at the open curtain, a man in a brown cardigan on the bed.

First-and-last-frame on Wan 3.0 locks where a long take starts and ends. So 20 seconds of camera movement still lands on your cliffhanger shot. Writing 'one continuous tracking shot, no cuts' stops the model from splitting it.

CC BY 4.0: use it however you like, just link back to OpenMicroDrama.