Rain streaks down the glass face of a black skyscraper at night, city lights blurred in the wet air far below. [Pedestal up] The camera rises slowly along the tower's edge past dark floors, then [Push in] toward the single lit penthouse floor at the top, where a lone figure in a charcoal suit stands at the window. Wet reflections shimmer, thin fog drifts between buildings. Cinematic, cool blue and amber tones, realistic rain.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 10s, 768P, prompt optimizer off
Two bracket commands in a row give Hailuo a clear rise-then-push path. Make the 9:16 first frame a low shot of the tower in rain.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Cufflink fastened before the signing
CEO romanceInsert
Extreme close-up of a man's wrist in a charcoal three-piece suit sleeve. His fingers fasten a silver cufflink with one precise twist, then his hand moves to a black fountain pen resting on a thick contract. [Static shot] then [Push in] slowly onto the pen nib as it touches the signature line. Warm desk-lamp light, shallow depth of field, rain-blurred city lights behind.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 6s, 1080P, prompt optimizer off
Hailuo handles hands best in inserts. One small action plus one camera move keeps fingers from warping. Cut it in right before Adrian speaks.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Midnight summons, over Adrian's shoulder
CEO romanceOver-the-shoulder
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical over-the-shoulder shot from behind a 32-year-old man with a sharp jaw, slicked-back black hair and cold dark eyes, wearing a charcoal three-piece suit with silver cufflinks, seated at a glass desk in a dark penthouse office; his shoulder and the back of his head fill the left foreground, slightly out of focus. In the sharp background, a 24-year-old woman with chestnut hair in a low ponytail, wearing a cream silk blouse and a thrifted tan trench coat stands in the doorway holding a dripping umbrella. The camera pushes in with small amplitude at slow speed toward her. The young woman with a clear, slightly nervous voice (S1) says: <d>[English] You wanted me here at midnight?</d> The man tilts his head a fraction and says nothing.
overall_soundscape: Heavy rain against floor-to-ceiling windows, the faint hum of the city far below, water dripping from the umbrella onto marble.
non_diegetic_music: A low sustained cello note with sparse, slow piano notes.
Settings: MiniMax H3 text-to-video, 9:16, 8s, 768P (regenerate keepers to 2K), native audio, prompt expansion disabled if your provider offers it
We describe the shot as foreground plus background, like H3's prompt guide does. Only one character gets a speaker ID, so Adrian stays silent. Draft at 768P, then redo the keeper at 2K.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Elevator stop: three-shot tension beat
CEO romanceClose-up
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical medium two-shot inside a mirrored private elevator at night. On the left stands a 32-year-old man with a sharp jaw, slicked-back black hair and cold dark eyes, wearing a charcoal three-piece suit with silver cufflinks; on the right stands a 24-year-old woman with chestnut hair in a low ponytail, wearing a cream silk blouse and a thrifted tan trench coat, clutching a folder. The elevator lurches, the lights cut out, and only a dim red emergency light remains. The camera holds a static shot. [Shot 2] At 00:05.000, the camera cuts to a close-up of the woman's face in red light, her breathing quickening as she glances sideways. The young woman with a soft, tense voice (S1) says: <d>[English] Is it stuck?</d> [Shot 3] At 00:10.000, the camera cuts to a close-up of the man; the camera pushes in with small amplitude at slow speed as he turns his head toward her. The man with a low, calm, unhurried voice (S2) says: <d>[English] Only if I want it to be.</d> His expression doesn't change.
overall_soundscape: The mechanical groan of the elevator stopping, a soft electrical buzz from the emergency light, the rustle of the folder in her hands, her quick breathing.
non_diegetic_music: A slow pulse of low synth bass with a single high violin note that rises in volume and cuts out on the final line.
Settings: MiniMax H3 text-to-video, 9:16, 15s, 768P then 2K regeneration, native audio, prompt expansion disabled if your provider offers it
Three timed shots in one 15-second clip keep both faces the same. You don't have to stitch. Each line gets its own close-up, so only one mouth moves.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Fake-engagement offer with locked cast references
CEO romanceMedium
References: Image 1 is Adrian Vale, a 32-year-old man with a sharp jaw, slicked-back black hair and cold dark eyes, wearing a charcoal three-piece suit with silver cufflinks. Image 2 is Elena Brooks, a 24-year-old woman with chestnut hair in a low ponytail, wearing a cream silk blouse and a thrifted tan trench coat. Image 3 is the glass penthouse office at night with a rain-streaked city view. Audio 1 is Adrian's voice timbre; Audio 2 is Elena's voice timbre. Keep both faces, outfits and the office layout exactly as referenced.
9:16 vertical, live-action, cinematic. [Shot 1] Medium shot: Adrian slides a black folder across the glass desk toward Elena, who stands opposite him. Adrian, in the voice of Audio 1, says: "One year as my fiancée. Then you walk away rich." Elena's lips stay closed; her eyes stay on the folder. [Shot 2] At 00:06.000, cut to a medium close-up of Elena. She opens the folder, reads, then looks up. Elena, in the voice of Audio 2, says: "And if I fall for you?" Adrian's lips stay closed. [Shot 3] At 00:10.000, cut to a medium shot of Adrian leaning back in his chair with the faintest smile. Sound: rain on glass, paper sliding, quiet city hum. Music: low strings at a slow tempo.
Naming each reference by number and job stops H3 from blending faces. Telling the listener to keep their lips closed stops shared mouth movement. Reuse one sheet per lead every episode.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Gala entrance, the room turns
RevengeWide
A 28-year-old woman with a sleek platinum bob, a black satin dress and red lipstick steps through tall double doors into a crowded, gold-lit ballroom. Guests in evening wear stop mid-conversation and turn toward her one by one; a champagne tray tilts in a waiter's hand. [Static shot] at first, then [Pull out] to reveal the whole room and the chandelier above. Cinematic, warm gold light, light haze, realistic fabric movement.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 6s, 1080P, prompt optimizer off
Spell out the order of the crowd's reaction, and Hailuo follows it. Pulling out shows the room's size, which a narrow 9:16 frame hides.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Courthouse steps through the camera flashes
RevengeTracking
A 28-year-old woman with a sleek platinum bob, a black satin dress and red lipstick walks down wide stone courthouse steps in daylight. Reporters crowd both sides, camera flashes pop and microphones push toward her; she doesn't slow down. [Tracking shot] from the front as she descends, then [Pan right] to catch a 35-year-old man in a navy suit with a smug half-smile, frozen at the top of the steps behind her. Natural daylight, crisp shadows, realistic crowd motion.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 10s, 768P, prompt optimizer off
Tracking from the front keeps her face in view the whole time. The late pan then hands the frame to Grant. Hailuo 2.3 only makes 10-second clips at 768P.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Vanessa's champagne starts to shake
RevengeClose-up
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical close-up of a 26-year-old woman in a pink tweed set with an oversized diamond ring, holding a champagne flute at a glittering gala. She is mid-smile when her eyes catch something off-screen; the smile freezes, the champagne trembles in the glass, and she lowers it slowly. The camera pushes in with small amplitude at slow speed toward her eyes.
overall_soundscape: The murmur of a crowded ballroom, a string quartet far away, the faint clink of her ring against the glass.
non_diegetic_music: A single low piano note held under a rising string tremolo.
H3 Max at 480P is a cheap way to test a reaction before paying for 2K. Shaking liquid shows fear better than a vague emotion word. Rerun the winner on H3.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Divorce papers over Grant's shoulder
RevengeOver-the-shoulder
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical over-the-shoulder shot from behind a 35-year-old man in a navy suit with a smug half-smile, seated at the head of a long candlelit dining table in a dim family mansion; his shoulder fills the right foreground. Across the table, a 28-year-old woman with a sleek platinum bob, a black satin dress and red lipstick stands and slides a stack of papers toward him with two fingers. The camera holds a static shot. The woman with a low, even, cold voice (S1) says: <d>[English] Sign it. You have until the candles burn out.</d> The man lets out a short laugh. The man with a smug, relaxed voice (S2) says: <d>[English] You'll be back.</d>
overall_soundscape: A grandfather clock ticking, candle flames fluttering, paper sliding across polished wood.
non_diegetic_music: A low cello ostinato at a slow tempo.
Settings: MiniMax H3 text-to-video, 9:16, 10s, 768P (regenerate keepers to 2K), native audio, prompt expansion disabled if your provider offers it
Give each speaker a voice note, cold against smug. Then H3 casts two different voices in one go. One sentence per line keeps 10 seconds from feeling rushed.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
The gala screen reveal (full beat)
RevengeInsert
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical insert close-up of a woman's hand with red nails pressing a small remote; behind it, out of focus, a giant screen hangs above a gala stage. [Shot 2] At 00:03.000, the camera cuts to a wide shot of the ballroom from the stage: the screen lights up with a blurred photo of a man in a navy suit kissing a woman in a pink tweed set, and guests gasp and turn. [Shot 3] At 00:07.000, the camera cuts to a medium shot of a 35-year-old man in a navy suit with a smug half-smile, his smile collapsing, beside a 26-year-old woman in a pink tweed set with an oversized diamond ring, who covers her face. [Shot 4] At 00:11.000, the camera cuts to a close-up of a 28-year-old woman with a sleek platinum bob, a black satin dress and red lipstick; the camera pushes in with small amplitude at slow speed as she raises her glass. The woman with a low, even, cold voice (S1) says: <d>[English] Happy anniversary, darling.</d>
overall_soundscape: A remote clicks, a projector hum swells, a wave of gasps and whispers rolls through the room, a glass shatters on the marble floor.
non_diegetic_music: Silence until the screen lights up, then a hard staccato string hit and a slow, heavy drum pulse under the final line.
Settings: MiniMax H3 text-to-video, 9:16, 15s, 768P then 2K regeneration, native audio, prompt expansion disabled if your provider offers it
Four timed shots tell the whole payoff in one clip. So the cast and ballroom match across every cut. Opening on the insert gives you the 3-second hook.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Silvercrest lodge under a full moon
WerewolfEstablishing
A huge full moon hangs over a misty pine forest at night; a long timber lodge with glowing windows sits in a clearing, wood smoke curling from its chimney. Fog rolls slowly between the trunks. [Pan right] slowly across the treeline, then [Push in] toward the lodge as dark shapes move between the trees. Cold blue moonlight against warm window light, cinematic, realistic fog.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 10s, 768P, prompt optimizer off
Hailuo does layered mood well when fog, smoke and moonlight each move. Use it as the cold open every episode so viewers know where they are.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Ivy runs through the pines
WerewolfTracking
A 22-year-old woman with freckles, auburn curls and an oversized red flannel shirt sprints through a dark pine forest at night, breath fogging, glancing back over her shoulder. Branches whip past and pine needles kick up under her boots. [Tracking shot,Shake] from the front at running speed. Moonlight through the trees, realistic hair and fabric motion.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 6s, 1080P, prompt optimizer off
Running is easy for Hailuo 2.3. [Tracking shot,Shake] gives you handheld panic in one command. Her red flannel stands out against dark trees on a phone.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Alpha's eyes flash amber
WerewolfClose-up
Close-up of a 30-year-old man with shoulder-length dark hair, a scar through his left eyebrow and a black leather jacket, standing in moonlight. He holds still, jaw tight, then his eyes slowly flash bright amber and his nostrils flare as if he caught a scent. [Static shot] then [Push in] slowly toward his eyes. Cold blue moonlight, shallow depth of field, faint breath fog.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 6s, 1080P, prompt optimizer off
Small face movements are Hailuo 2.3's strength. Keep the action tiny and in order: still, flash, flare. Light the eyebrow scar in your first frame, since it marks him.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
The rejection ceremony (full beat)
WerewolfWide
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical wide shot of a moonlit forest clearing where dozens of pack members stand in a circle around a bonfire under a huge full moon. In the center, a 30-year-old man with shoulder-length dark hair, a scar through his left eyebrow and a black leather jacket faces a 22-year-old woman with freckles, auburn curls and an oversized red flannel shirt. The camera pedestals down with small amplitude at slow speed. [Shot 2] At 00:05.000, the camera cuts to a medium shot of the man, firelight on one side of his face, his eyes flashing amber. The man with a deep, rough, steady voice (S1) says: <d>[English] I, Kade Thorne, reject you as my mate.</d> [Shot 3] At 00:10.000, the camera cuts to a close-up of the woman; tears stand in her eyes, but her chin lifts. The young woman with a shaky but defiant voice (S2) says: <d>[English] Then I reject you first.</d> Behind her, the crowd murmurs.
overall_soundscape: A crackling bonfire, wind moving through the pines, a distant wolf howl, and a ripple of shocked murmurs after the last line.
non_diegetic_music: Low taiko drum hits at a slow tempo under sustained choir tones that swell, then drop to silence after the final word.
Settings: MiniMax H3 text-to-video, 9:16, 15s, 768P then 2K regeneration, native audio, prompt expansion disabled if your provider offers it
H3 keeps the pack, the firelight and both leads steady across all three cuts. Each line gets its own shot, so only one mouth moves. Her comeback is your hook.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Lodge doorway face-off with cast references
WerewolfOver-the-shoulder
References: Image 1 is Kade Thorne, 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 Hart, a 22-year-old woman with freckles, auburn curls and an oversized red flannel shirt. Image 3 is the timber pack lodge porch at night. Audio 1 is Kade's voice timbre; Audio 2 is Ivy's. Keep both faces, outfits and the porch exactly as referenced.
9:16 vertical, live-action, cinematic, night, light snowfall. [Shot 1] Over-the-shoulder from behind Ivy on the porch steps, looking up at Kade in the lit doorway. Kade, in the voice of Audio 1, says: "You shouldn't have come back." Ivy's lips stay closed. [Shot 2] At 00:06.000, cut to a reverse over-the-shoulder from behind Kade, looking down at Ivy. Ivy, in the voice of Audio 2, says: "I didn't come back for you." Kade's lips stay closed as his eyes flash amber. Sound: wind, creaking boards, a distant howl. Music: a low drone at a slow tempo.
Most dialogue scenes run on reverse over-the-shoulder pairs. References stop the two angles from becoming two people. Shoot both sheets in the season costume.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Carriage arrives at the rain-soaked estate
HistoricalEstablishing
A grand Regency stone manor at dusk in heavy rain, lantern light glowing at the entrance. A black horse-drawn carriage rolls up the gravel drive, wheels throwing water, the horses' breath steaming. [Pedestal down] from the rooftops to the drive, then [Push in] as a footman with an umbrella hurries to the carriage door. Muted gray-blue palette with warm lantern accents, period-accurate, cinematic.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 10s, 768P, prompt optimizer off
Horses and turning wheels mean lots of motion, which suits Hailuo. Starting high shows the house's height. The footman's dash gives you a clean cut.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Eleanor's curtsy in the candlelit ballroom
HistoricalMedium
Medium shot of a 20-year-old woman with dark curls pinned up, a pale blue empire-waist gown and pearl drop earrings in a candlelit Regency ballroom. She sinks into a slow, graceful curtsy with her eyes down, then lifts her gaze sharply toward someone off-screen and snaps her fan shut. [Truck left] slowly, keeping her centered. Warm candlelight, soft haze, dancers blurred in the background.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 6s, 1080P, prompt optimizer off
One smooth gesture (curtsy, look, fan snap) suits Hailuo. The snap is a natural cut. A slow truck (sideways camera slide) adds life without fighting the action.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Breaking the duke's wax seal
HistoricalInsert
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical insert close-up of a woman's gloved hands holding a folded letter sealed with dark red wax stamped with a raven crest, lit by a single candle. Her thumb presses the seal until it cracks; she unfolds the letter and the candle flame wavers. The camera holds a static shot.
overall_soundscape: The sharp crack of wax, paper crinkling as it unfolds, a soft candle flicker, rain against a window.
non_diegetic_music: N/A
Inserts don't need dialogue, so H3 Max is enough. You get the wax crack sound for free. Setting music to N/A keeps your score track free for the scene.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
The duke's ultimatum in the carriage
HistoricalClose-up
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical close-up of a 29-year-old man with cold gray eyes, wearing a black tailcoat and a white cravat, seated in a dim carriage interior; rain streaks the small window beside him and a hanging lantern sways, moving light across his face. The camera pushes in with small amplitude at slow speed. He looks straight at someone opposite him. The man with a quiet, clipped, aristocratic English voice (S1) says: <d>[English] Marry me by Sunday, Lady Ashford, or your father's debts become very public.</d> He leans back into shadow.
overall_soundscape: Carriage wheels on wet cobblestones, rain drumming on the roof, horses' hooves, the lantern creaking on its hook.
non_diegetic_music: A low harpsichord figure at a slow tempo over a sustained cello.
Settings: MiniMax H3 text-to-video, 9:16, 10s, 768P (regenerate keepers to 2K), native audio, prompt expansion disabled if your provider offers it
A swaying lantern gives you moving light on a still face. Accent and pace shape the delivery more than an emotion word. Ten seconds fits one long line plus a pause.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Waltz in three chained clips
HistoricalTracking
Clip A (first frame: the couple at the edge of the dance floor): A man takes the hand of a young woman and leads her onto a candlelit ballroom floor. He is a 29-year-old man with cold gray eyes, wearing a black tailcoat and a white cravat; she is a 20-year-old woman with dark curls pinned up, a pale blue empire-waist gown and pearl drop earrings. [Tracking shot] following them as other couples part.
Clip B (first frame: the last frame of Clip A): The couple starts a slow waltz, his hand at her back, her gown swirling. [Truck right,Pan left] circling them as candlelight flickers across their faces; she refuses to meet his eyes.
Clip C (first frame: the last frame of Clip B): Mid-turn, he pulls her closer and leans to her ear. She stops dancing and stares at him, stunned, while the other couples keep spinning around them. [Push in] slowly toward her face, then [Static shot].
Settings: Hailuo 2.3 image-to-video, run 3 times, 9:16 first frames, 6s each, 1080P, prompt optimizer off
Hailuo 2.3 has no last-frame input. So save each clip's final frame and use it as the next first frame. Describe each clip from its own start pose so the dance doesn't restart. Run Clip A, B and C with the same settings, then cut on the turns.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Lighthouse beam over the drowned town
ThrillerEstablishing
A rain-soaked coastal town at night seen from a cliff: wet rooftops, one flickering neon motel sign, and a lighthouse whose beam sweeps slowly across the harbor and the town. Waves smash against the rocks below. [Static shot] for the first sweep of the beam, then [Zoom in] slowly toward the motel as one room's light switches off. Desaturated teal and sodium orange, heavy rain, cinematic.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 10s, 768P, prompt optimizer off
A still camera lets the lighthouse beam do the moving. The light switching off is a clue viewers remember. Ten slow seconds suit a cold open.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Flashlight down the motel walkway
ThrillerTracking
A detective moves down a narrow motel exterior walkway at night: a 34-year-old woman with short black hair and tired eyes, wearing an olive raincoat, flashlight raised, her other hand resting on her holster. Rain pours off the roof edge; doors slide past, one of them ajar. [Tracking shot,Shake] following just behind her shoulder, until she stops dead at the open door. Flickering neon, wet concrete reflections, a realistic flashlight beam.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 6s, 1080P, prompt optimizer off
A shaky follow shot keeps her face mostly hidden, so small drift won't show. Use it as filler between dialogue shots, and cut on the stop.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Polaroid of Nora, taken from behind
ThrillerInsert
Insert close-up of a Polaroid photo lying face-up on a wet motel carpet, lit by a flashlight beam. The image slowly develops from blank into a blurry photo of a woman with short black hair in an olive raincoat, taken from behind. A drop of rain lands on it. [Static shot] then [Push in] slowly onto the photo. High contrast, cold light.
Settings: Hailuo 2.3 Fast image-to-video, 9:16 first frame, 6s, 768P
It's a cheap insert, so Hailuo 2.3 Fast is plenty. The slow reveal works as a silent cliffhanger. Plant it 60 to 90 seconds in to end an episode.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Interrogation across the steel table
ThrillerOver-the-shoulder
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical over-the-shoulder shot from behind a 34-year-old woman with short black hair and tired eyes, wearing an olive raincoat, seated at a steel table in a small interrogation room; her shoulder fills the left foreground. Across from her sits a 31-year-old man with wire-rimmed glasses and a gray cardigan, hands folded, perfectly calm under a buzzing fluorescent light. The camera pushes in with small amplitude at slow speed toward him. The woman with a tired, flat, low voice (S1) says: <d>[English] Where were you at two a.m., Eli?</d> The man smiles politely and waits a beat. The soft-spoken man with a gentle, unhurried voice (S2) says: <d>[English] Watching you, Detective.</d>
overall_soundscape: A fluorescent tube buzzing, the scrape of a chair leg, rain faintly hitting a high window, a long silence after the last line.
non_diegetic_music: A low, slow synth drone that drops out completely on his final word.
Settings: MiniMax H3 text-to-video, 9:16, 12s, 768P (regenerate keepers to 2K), native audio, prompt expansion disabled if your provider offers it
Opposite voice notes make the polite threat land. The written pause gives H3 room to time the reply. Cutting the music on his last word still works.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
The call in the parked car (full beat)
ThrillerMedium
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical medium shot through the windshield of a parked car at night in heavy rain; inside sits a 34-year-old woman with short black hair and tired eyes, wearing an olive raincoat. She answers a buzzing phone and presses it to her ear. The camera holds a static shot. A distorted, calm male voice on the phone (S1) says in an off-screen voiceover: <d>[English] You're looking for the wrong man, Nora.</d> while her lips remain completely closed. [Shot 2] At 00:06.000, the camera cuts to a close-up of her face lit by the phone screen; her eyes move to the rearview mirror. The woman with a tired, flat, low voice (S2) says: <d>[English] Who is this?</d> [Shot 3] At 00:10.000, the camera cuts to an insert of the rearview mirror: in the back seat, a man's silhouette with the glint of wire-rimmed glasses sits perfectly still. The camera pushes in with small amplitude at fast speed.
overall_soundscape: Rain hammering on the car roof, a phone buzzing, the wipers squeaking once, then sudden silence when the mirror reveal hits.
non_diegetic_music: A ticking hi-hat pattern at a moderate tempo with a low synth that swells and cuts off at the mirror reveal.
Settings: MiniMax H3 text-to-video, 9:16, 15s, 768P then 2K regeneration, native audio, prompt expansion disabled if your provider offers it
MiniMax's off-screen voiceover wording keeps Nora's mouth still while the caller talks. Then she gets her own speaker ID. The mirror insert is a ready-made paywall cliffhanger.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Food truck smoke clears the wedding lawn
ComedyWide
Wide shot of a sunny garden wedding lawn with white chairs and a floral arch. A battered food truck parked at the edge suddenly belches a thick cloud of gray smoke from its roof vent. Guests in pastel outfits leap up and scatter, hats flying, while a flower girl keeps eating a taco, unbothered. [Pull out] to reveal the full chaos. Bright daylight, vivid colors, comedic timing, realistic smoke.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 6s, 1080P, prompt optimizer off
Physical comedy reads best wide. One calm character in the chaos, the flower girl, is the punchline. Hailuo's smoke and crowd motion carry the gag.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Tortilla flip lands on the clipboard
ComedyMedium
Medium shot through a cramped food truck window: a 27-year-old man with curly hair and a mustard-stained apron flips a tortilla high off the griddle with a showy flourish. It sails out the window and lands flat on the clipboard of a 26-year-old woman with a headset and a color-coded clipboard, who is standing outside. She freezes and stares at it. [Static shot] the whole time. Bright daylight, saturated colors, comedic timing.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 6s, 1080P, prompt optimizer off
A locked camera lets the object's path be the joke. Say where the tortilla starts and lands, so Hailuo commits to the arc.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Priya's eye twitch
ComedyClose-up
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical close-up of a 26-year-old woman with a headset and a color-coded clipboard, with a tortilla stuck to the clipboard, in bright daylight at an outdoor wedding. She stares straight into the lens, completely deadpan; one eyelid twitches twice and her headset crackles. The camera holds a static shot.
overall_soundscape: Wedding guests chattering nearby, a headset crackle, a distant sizzle from a griddle.
non_diegetic_music: A single plucked cello note, then silence.
Reaction close-ups make up most of a comedy edit. H3 Max at 480P keeps them cheap. The one-note music cue works like a rimshot.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Food truck vs. planner: the first argument
ComedyOver-the-shoulder
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical over-the-shoulder shot from behind a 26-year-old woman with a headset and a color-coded clipboard, standing outside a battered food truck at a garden wedding; her shoulder and headset fill the right foreground. In the truck window, a 27-year-old man with curly hair and a mustard-stained apron leans out, grinning. The camera holds a static shot. The woman with a fast, clipped, exasperated voice (S1) says: <d>[English] You're the caterer?</d> The man shrugs. The man with a warm, easygoing voice (S2) says: <d>[English] I'm catering-adjacent.</d>
overall_soundscape: A griddle sizzling, wedding guests chatting, a bee buzzing past, a string quartet warming up somewhere.
non_diegetic_music: A bouncy pizzicato string line at a quick tempo.
Settings: MiniMax H3 text-to-video, 9:16, 8s, 768P (regenerate keepers to 2K), native audio, prompt expansion disabled if your provider offers it
Keep the punchline shorter than the setup so the joke can land. Opposite voice notes do half the character work before anyone moves.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
The wedding cake run (full beat)
ComedyTracking
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical tracking shot in front of a 27-year-old man with curly hair and a mustard-stained apron as he hurries across a garden wedding lawn carrying a wobbling three-tier white cake. The camera tracks backward ahead of him at fast speed. [Shot 2] At 00:04.000, the camera cuts to a medium shot of a 26-year-old woman with a headset and a color-coded clipboard, sprinting after him and waving the clipboard. The woman with a fast, clipped, exasperated voice (S1) shouts: <d>[English] Benny, not across the grass!</d> [Shot 3] At 00:08.000, the camera cuts to a wide shot as his foot catches a sprinkler; he stumbles and spins but keeps the cake upright as the sprinkler soaks the front row of guests. [Shot 4] At 00:12.000, the camera cuts to a close-up of the man, dripping wet, holding the perfect cake. The man with a warm, easygoing voice (S2) says: <d>[English] Nailed it.</d>
overall_soundscape: Fast footsteps on grass, the cake board creaking, a sprinkler hissing on, guests shrieking and laughing.
non_diegetic_music: An upbeat brass and snare march at a fast tempo that stops dead on the final line.
Settings: MiniMax H3 text-to-video, 9:16, 15s, 768P then 2K regeneration, native audio, prompt expansion disabled if your provider offers it
Four quick cuts in 15 seconds give the gag comedy pacing. Keeping the cake in every shot helps H3 hold continuity. A two-word last line lets the music stop dead.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Floating citadel at dawn
FantasyEstablishing
A vast white-stone citadel floats above a sea of clouds at dawn, broken rock and waterfalls hanging beneath it and pouring into the mist. Birds circle its towers; a narrow crystal bridge glints in the sunrise. [Pedestal up] from below the cloud layer to reveal the citadel, then [Push in] toward the highest tower. Golden-pink light, epic scale, painterly cinematic realism.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 10s, 768P, prompt optimizer off
Hailuo 2.3 handles big effects shots with falling water and drifting rock. Rising through clouds is a strong reveal in a tall frame. Reuse it as the location card.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
The forbidden rune wakes
FantasyInsert
Extreme close-up of a young woman's right wrist with a faint blue rune tattoo, the frayed edge of a worn moss-green cloak sleeve at the side of the frame. The rune begins to glow, and lines of blue light crawl outward across her skin like frost as her fingers curl into a fist. [Push in] slowly onto the rune. Dark background, the glow lighting the cloak fabric, drifting magical particles.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 6s, 1080P, prompt optimizer off
Glows and particles play to Hailuo's strengths. A recurring mark helps viewers track who's who. Frame the wrist the same way every time.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Walking the crystal forest
FantasyWide
Wide shot of a forest of towering translucent crystal trees at twilight, faint blue and violet light pulsing inside them. A small figure with a silver-white braid and a worn moss-green cloak walks a path of glowing moss; each crystal she passes brightens, then dims. [Truck left] slowly alongside her. Low ground mist, soft bloom, fantasy cinematic.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 10s, 768P, prompt optimizer off
Tying the light change to her footsteps gives Hailuo a clear cause and effect. Her cloak color makes a small figure easy to spot in a wide shot.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
That mark is forbidden
FantasyMedium
integrated_multimodal_description: [Shot 1] Live-action, cinematic, high fantasy, a vertical medium shot in a moonlit crystal forest. In the path stands a 25-year-old man in dark engraved armor with a thin gold circlet and storm-gray eyes, blocking 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. She hides the wrist behind her back. The camera trucks right with small amplitude at slow speed. The man with a deep, measured, commanding voice (S1) says: <d>[English] That mark is forbidden.</d> The young woman with a bright, quick, defiant voice (S2) says: <d>[English] Then arrest me, Your Highness.</d>
overall_soundscape: Crystals chiming faintly in the wind, armor plates shifting, glowing moss rustling underfoot.
non_diegetic_music: A low French horn over soft harp arpeggios at a slow tempo.
Settings: MiniMax H3 text-to-video, 9:16, 10s, 768P (regenerate keepers to 2K), native audio, prompt expansion disabled if your provider offers it
A slow truck (sideways camera slide) keeps two standing characters from feeling static. Full descriptions let H3 keep both leads apart without references. It's a good test before you build character sheets.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Throne room standoff with cast and camera references
FantasyOver-the-shoulder
References: Image 1 is Sera Nightingale, 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 2 is Prince Caelan of Eryndor, a 25-year-old man in dark engraved armor with a thin gold circlet and storm-gray eyes. Image 3 is the throne room of the floating citadel. Video 1 provides the camera movement only: a slow arc around two people. Keep both faces, costumes and the room exactly as referenced.
9:16 vertical, live-action, high fantasy. [Shot 1] Over-the-shoulder from behind Caelan, seated on a white stone throne, looking down the steps at Sera, who stands bound in chains of light. Follow the slow arc from Video 1. Caelan, in a deep, measured, commanding voice, says: "Tell them you lied about the rune." Sera's lips stay closed. [Shot 2] At 00:07.000, cut to a reverse over-the-shoulder from behind Sera; her rune flares and the chains of light shatter into sparks. Sera, in a bright, quick, defiant voice, says: "I don't lie, Your Highness. I burn." Caelan's lips stay closed. Sound: chains humming, crackling magic, an echoing hall. Music: a rising choir over low brass.
Give the reference video one job, camera only, so H3 doesn't copy its people. Name each reference by role, as MiniMax's guide asks. Reference clips must run 2 to 15 seconds.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
The porch light comes on
Family dramaEstablishing
A small white clapboard house at dusk on a quiet suburban street, autumn leaves on the steps. The porch light flickers on. At the bottom of the steps, a 70-year-old man in a brown cardigan leaning on a walking cane looks up at the front door and doesn't move. [Static shot] then [Push in] slowly toward him. Warm porch light against a blue dusk sky, a light wind moving the leaves.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 10s, 768P, prompt optimizer off
Holding still before the push lets the porch light tell the story. Keeping Walter far away saves his face for a later close-up. It also works as a final image.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Grace takes the call on the night shift
Family dramaTracking
A 45-year-old woman with a graying bun and navy scrubs strides fast down a bright hospital corridor at night, phone pressed to her ear, dodging a gurney and a nurse with a cart. Her face shifts from focused to stunned as she listens, and she stops. [Tracking shot] in front of her, slightly low. Cool fluorescent light, realistic motion, shallow depth of field.
Settings: Hailuo 2.3 image-to-video, 9:16 first frame, 6s, 1080P, prompt optimizer off
Tracking from the front keeps her face readable. The mid-clip stop gives Hailuo time to show her face change. Add the phone audio in your edit, since Hailuo 2.3 is silent.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
The photo in Walter's hand
Family dramaInsert
Insert close-up of an old man's weathered hand holding a faded color photograph of a young woman in a graduation gown, the cuff of a brown cardigan sleeve visible. His thumb slowly brushes across her face in the photo, then his hand trembles slightly. [Static shot] then [Push in] slowly. Warm lamp light, shallow depth of field, dust motes in the air.
Settings: Hailuo 2.3 Fast image-to-video, 9:16 first frame, 6s, 768P
Small hand motion on a held object is a reliable Hailuo insert. 2.3 Fast is cheap enough for several takes. This shot does backstory work so no line has to.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Who is he, Mom?
Family dramaClose-up
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical close-up of a 16-year-old girl in an oversized hoodie with braces and headphones around her neck, sitting at a small kitchen table at night under a single pendant lamp. She looks past the camera toward the window, then back. The camera pushes in with small amplitude at slow speed. The teenage girl with a quiet, wary voice (S1) says: <d>[English] Mom, there's an old man on our porch. Who is he?</d>
overall_soundscape: A refrigerator hum, a clock ticking, wind against the window, then a muffled knock at the front door right after her line.
non_diegetic_music: Soft felt piano notes at a slow tempo.
Settings: MiniMax H3 text-to-video, 9:16, 8s, 768P (regenerate keepers to 2K), native audio, prompt expansion disabled if your provider offers it
The knock comes after her line, so it bridges into the next shot. A teen voice plus 'quiet, wary' is enough for H3 to cast her.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.
Twenty years on the porch (full beat)
Family dramaWide
integrated_multimodal_description: [Shot 1] Live-action, cinematic, a vertical wide shot of a small white clapboard house's porch at dusk. The front door opens and a 45-year-old woman with a graying bun and navy scrubs steps out, then stops. At the bottom of the steps stands a 70-year-old man in a brown cardigan leaning on a walking cane. The camera holds a static shot. [Shot 2] At 00:05.000, the camera cuts to a close-up of the woman; her jaw tightens. The woman with a tired, low, controlled voice (S1) says: <d>[English] Twenty years, Dad.</d> [Shot 3] At 00:09.000, the camera cuts to a close-up of the old man; his eyes are wet. The old man with a gravelly, soft voice (S2) says: <d>[English] I know. I counted every one.</d> [Shot 4] At 00:13.000, the camera cuts to a medium shot through the doorway behind the woman, where a 16-year-old girl in an oversized hoodie with braces and headphones around her neck watches from the hallway.
overall_soundscape: Wind through dry leaves, a screen door creaking, distant traffic, a dog barking far down the street.
non_diegetic_music: A single acoustic guitar with slow fingerpicking over a soft sustained string pad.
Settings: MiniMax H3 text-to-video, 9:16, 15s, 768P then 2K regeneration, native audio, prompt expansion disabled if your provider offers it
Family drama runs on close-ups and silence, so each line gets its own face. The last shot goes to the person with the most to lose. Lily watching sets up the next episode.
CC BY 4.0: use it however you like, just link back to OpenMicroDrama.