Alici.AI
Home/Create/AI Short Drama Generator

AI Short Drama Generator

Upload a script or a reference character, drop one prompt — Alici VideoAgent's Micro Drama Mode returns a polished vertical drama with voiceover, captions, and a ReelShort-ready cliffhanger.

Describe what you want to create...
+
Generate
Free to start · Sign in to generate
📋Prompt Recipe
Use This
# INPUT KIT — what the user feeds into Alici VideoAgent Micro Drama Mode
# (a) One reference photo of the female lead (clean headshot, any ethnicity)
# (b) One paragraph of script or situation — OR the full MASTER PROMPT below
# (c) Press Generate once
# That's it. VideoAgent returns a polished vertical drama from these three inputs alone.

# REFERENCE FORMULA
# Based on the cold-open formula popularized by "The Double Life of My Billionaire Husband"
# (ReelShort, 500 million+ views, the most-cited microdrama opener in NPR / Rolling Stone /
# Rest of World / EqualOcean coverage). The Secret Billionaire Reveal is the genre's #1
# eight-second hook — a woman dismisses a man she has decided is beneath her, then discovers
# he is secretly the wealthiest heir in the city. This prompt reproduces that beat as an
# Alici VideoAgent storyboard, optimized for the new Micro Drama Mode (vertical 9:16,
# eight-second hook, Seedance 2.0 / Sora 2 / Veo 3.1 / Hailuo 02 routing).

GLOBAL LOCK
Style: photoreal cinematic vertical microdrama, 9:16, 1080×1920, 24fps, shallow depth of field, warm cinematic grade with cool teal shadow rolloff, subtle film grain. Subject identity, wardrobe, and diner environment must remain locked across every shot. Two on-camera figures and only two: a woman in her mid-20s (any ethnicity, model default; shoulder-length wavy chestnut hair parted slightly off-center; faded denim jacket over a plain white tee; small silver hoop earrings; light blue jeans; tired but proud expression; emotional arc from dismissive to stunned) and a man in his early 30s (any ethnicity, model default; close-cropped hair; worn brown leather jacket over a plain dark t-shirt; faded jeans; scuffed brown leather work boots; calloused hands with a thin silver ring on the right pinky; carries a slim cream document folder in his right hand; expression neutral and unreadable throughout). Environment: a classic American roadside diner just before sunset, twenty miles outside a small Texas town. Long formica counter with a polished chrome rim. Cracked red vinyl swivel stools, four visible. A half-empty white ceramic coffee mug on a paper placemat in front of the woman, steam barely visible. A vintage Wurlitzer jukebox glowing red and amber against the back wall. A neon sign reading OPEN 24 HOURS hums above the kitchen pass-through, casting a faint pink wash on the right edge of frame. Through the front window: a gravel parking lot, a single dust-covered red Ford F-150 pickup with one bumper sticker, a US Route 66 sign post leaning slightly, golden hour light flooding low across the asphalt. Lighting: warm golden-hour key from the front window raking left-to-right at 15 degrees above horizon, cool fluorescent fill from the strip lights above the counter, motivated practical from the OPEN 24 HOURS neon on the right and the jukebox glow on the back wall. Camera language: 50mm full-frame look, gentle handheld micro-jitter, smooth dolly and slow push-ins only, no whip pans, no zoom blur, no rack focus. Speech: minimal dialogue, intimate close-mic ADR feel, no room reverb, English with neutral North American accent, lip-sync strictness high, but the only spoken line stays off-camera. Audio bed: a single low cello drone, no melody, swelling on the final shot. The jukebox plays nothing — it is on but silent.

[00:00–00:02] EXTREME WIDE establishing shot through the diner front window from the gravel parking lot. The dust-covered red Ford F-150 dominates the foreground out of focus; through the glass beyond it the woman sits alone at the counter scrolling her phone, the OPEN 24 HOURS neon glowing pink behind her, the Route 66 sign post visible at the right edge of frame. Slow 5% dolly-in. No dialogue. Only the cello drone, distant cicadas, and the faint hum of the neon. Wardrobe, F-150, and diner details must match the GLOBAL LOCK exactly.

[00:02–00:04] CUT to a MEDIUM SHOT inside the diner from the woman's eyeline. The man slides onto the red vinyl stool beside her, places the cream document folder face-down on the formica counter, says nothing, does not look at her. The woman glances at him with mild irritation, takes in the worn leather jacket, the scuffed boots, and the calloused hand, smirks faintly, decides he is a drifter passing through, and looks back at her phone. Camera holds static. The cello drone gains a single low note. The jukebox glow flickers once on his right shoulder.

[00:04–00:06] CLOSE-UP, eye-level, the man's calloused hand sliding the cream folder across the counter toward the woman. He flips it open with a single thumb. Inside, the top page is a property deed, beneath it a mortgage payoff certificate stamped PAID IN FULL, and resting on top a heavy brass key on a leather fob embossed with a single letter — no other readable text anywhere in the folder. His voice, off-camera, low and even with a faint Texas drawl: "It was always in your name. He just never told you." Lip-sync not required. Subtle handheld micro-jitter. The cello drone pitches up a half-step. The neon sign flares slightly on the right edge of frame.

[00:06–00:08] MEDIUM CLOSE-UP, slow 8% push-in onto the woman's face as her eyes scan the deed. Her smirk falls. Her breath catches. She does not turn her head. Her lips part but no sound comes out. The OPEN 24 HOURS neon behind her softens into pink bokeh, the jukebox into amber. Cut to black on the eighth second exactly as the cello swell peaks — the cliffhanger lands on her stunned silent face, not on a spoken word.

NEGATIVE PROMPT
no extra figures or background extras inside the diner, no waitress, no cook in the back, no other patrons, no warped hands or fingers, no readable text on the documents beyond the embossed single letter on the keyring and the PAID IN FULL stamp, no jewelry changes between shots, no wardrobe color drift, no morphing facial features between cuts, no lip-sync mismatch on the off-camera line, no smile or laugh on the woman, no fast camera moves or whip pans, no jukebox music — silent jukebox only, no neon glare flicker beyond the one scripted flare, no over-saturated grade, no crushed blacks, no rolling-shutter wobble, no obvious AI motion blur, no plastic skin, no robotic voice cadence on the off-camera line, no music with melody — drone only, no extra dialogue, no on-screen captions, no watermark, no celebrity likeness, no Asian skyline cues, no European architecture, no urban setting.

DELTA PROMPT — top drift risks and corrective micro-prompts
1. Hair length drift → append "shoulder-length wavy chestnut hair, parted slightly off-center, no bangs, length unchanged across all shots".
2. Wardrobe color shifts → append "faded blue denim jacket and plain white tee locked, no black, no gray, no plaid".
3. Extra background characters → append "only two figures inside the diner, four empty stools otherwise, no waitress, no cook, no other patrons".
4. Documents acquire readable text → append "documents are blank cream cardstock except for one PAID IN FULL stamp and the embossed single letter on the brass keyring, no other printed words anywhere".
5. Man delivers an on-camera line → append "man speaks only off-camera at [00:04–00:06], remains silent on-camera, lips do not move when the line is heard".
6. Cliffhanger cuts on a word → append "final cut lands at exactly 00:08:00 on the cello peak, not on dialogue, on the woman's silent stunned face".
7. Diner becomes a coffee shop, bar, or city restaurant → append "1950s American roadside diner only, formica counter, red vinyl stools, neon OPEN 24 HOURS sign, vintage Wurlitzer jukebox, no espresso machine, no liquor bottles, no urban window view".
8. Lighting drifts to noon or night → append "golden hour key light only, sun angle 15 degrees above horizon, warm 3200K, neon and jukebox as motivated fill, no overhead noon sun".
9. Music acquires melody → append "single sustained low cello drone, no melodic line, no percussion, no piano, jukebox is silent, swell only on the final two seconds".
10. Camera adds whip pans, zooms, or rack focus → append "all moves are dolly or static, micro-handheld jitter only, no whip pan, no zoom blur, no rack focus, the F-150 stays out of focus throughout the establishing wide".

# POLISHED OUTPUT KIT — what Alici VideoAgent Micro Drama Mode returns in ONE generate call
# - Vertical 9:16 1080×1920 24fps MP4
# - Off-camera voiceover (TTS, neutral North American accent with faint Texas drawl)
# - Burned-in English captions
# - Cello drone audio bed (no melody, swell on the cliffhanger)
# - Four locked shots over eight seconds
# - Character consistency guarantee: reuse this GLOBAL LOCK verbatim for episodes 2-5
#   and the same face, wardrobe, diner, and F-150 will carry forward
# - Ready to upload to ReelShort, DramaBox, TikTok PineDrama, or Instagram Reels as-is

Why One Prompt Can Become a Polished Billionaire Reveal

The most-watched scene in the history of vertical microdrama is a status reversal that lasts eight seconds. The Double Life of My Billionaire Husband crossed five hundred million views on ReelShort by opening with a woman at a worn formica counter dismissing a man in a leather jacket — only to discover he is secretly the wealthiest heir in the city. That eight-second beat earned ReelShort a one-point-two billion dollar business, and it is the scene NPR, Rolling Stone, Rest of World, and EqualOcean all reach for when they need to explain microdrama to a new reader.

The reason one prompt can reproduce it is structural. The secret billionaire reveal only needs two locked characters, a static interior, one off-camera line, and a slow push-in on a stunned face. Those are the four conditions where Seedance 2.0, Sora 2, Veo 3.1, and Hailuo 02 all stop hallucinating — no crowds to render, no chase scenes, no lip-sync to match. The genre's most-paid-for moment also happens to be the easiest thing an AI video model can get right today.

Alici VideoAgent's Micro Drama Mode bundles all of that into a single Generate call: one reference photo of the lead, one prompt describing the beat, and the system returns a polished vertical cut with voiceover, captions, and the cello cliffhanger baked in. No editing timeline, no voiceover session, no separate subtitle job. One prompt, one output — ready to upload to ReelShort, DramaBox, or PineDrama.

Key Insight: In microdrama, the shortest scene pays the most — and the shortest scene is also the only kind AI video models render without tells. One prompt is enough because the genre only asks for eight seconds.

How One Prompt Becomes a Polished Drama in Micro Drama Mode

Tip 1: Upload one reference photo of your lead

Micro Drama Mode uses the face as a character lock across all four shots. Upload a single clean headshot and the woman in the establishing wide is guaranteed to be the same woman in the closing push-in — the one feature Pippit, Hailuo, and microdrama.ai do not guarantee today.

Change: upload a single clean headshot before pasting the prompt

Tip 2: Paste one prompt or a one-paragraph script

The MASTER PROMPT above is the full template for a secret billionaire reveal, but you can replace it with a single paragraph like: "She discovers her quiet neighbor is secretly the CEO of the company that just fired her." Micro Drama Mode builds the four-shot GLOBAL LOCK, the off-camera line, and the cliffhanger timing automatically.

Change: paste the REFERENCE FORMULA plus GLOBAL LOCK, or a one-paragraph situation, into the prompt field

Tip 3: Hit Generate once — get a polished cut back

No separate voiceover session. No separate captions pass. No manual 9:16 crop. Micro Drama Mode routes the prompt through Seedance 2.0 for character consistency, Sora 2 for the dolly push-in, Veo 3.1 for the off-camera voice, and bundles auto captions plus the cello audio bed into a single 1080×1920 MP4.

Change: press Generate once — do not open a second tool

Common Pitfall: Creators paste a multi-scene novel expecting a full episode. Micro Drama Mode is tuned for the eight-second cliffhanger hook, not a sixty-episode arc. Give it one beat, one reveal, one face — and let the cliffhanger land. You build the series by reusing the same GLOBAL LOCK across calls, one episode at a time.

Frequently Asked Questions

What exactly do I need to give Alici VideoAgent — a full script or just a prompt?

One reference photo of your lead plus one prompt is enough. Micro Drama Mode routes the prompt through Seedance 2.0, Sora 2, Veo 3.1, or Hailuo 02 and returns a polished eight-second cut with voiceover, captions, and the audio bed included. You do not need a shot list, a voiceover file, a subtitle track, or an editing timeline.

Does the same character show up in episodes two, three, and four?

Yes, as long as you reuse the identical GLOBAL LOCK paragraph across calls. Micro Drama Mode locks identity on Seedance 2.0's character-consistency layer, so the same woman, same diner, same Ford F-150, and same lighting carry into episode two when you append a new four-shot timeline under the same lock. This is the one feature that Pippit, Hailuo, and microdrama.ai do not guarantee today.

What does "polished" actually include, and where can I publish it?

Every Micro Drama Mode output is a 9:16 vertical 1080×1920 24fps MP4 with off-camera voiceover, burned-in English captions, and the audio bed baked into the track. ReelShort, DramaBox, TikTok PineDrama, and Instagram Reels all accept the file as-is — the polished eight-second reveal becomes the cold open of your sixty-to-ninety-second episode one, delivered straight to the paywall.