Duanju Generator for Vertical Microdramas

A duanju (短剧) tells a short story in quick episodes. Start with one moment: a discovery, a reaction or a cliffhanger. Pick a shot from the storyboard below or write your own, then generate a vertical clip.

credits : 1

Storyboard: The Coat Check

These are three editable prompts for three separate clips. Choose a shot below the prompt field to load it.

The discovery

Vertical close-up in a quiet coatroom. A woman in a mustard coat reaches into her pocket and pulls out a small brass key tied with blue thread. Hold on the key as her fingers stop moving.

The reaction

Vertical medium close-up in the same quiet coatroom. The woman in the mustard coat lifts the brass key tied with blue thread. Behind the counter, an elderly attendant sees it and slowly takes off his glasses.

The cliffhanger

Vertical shot over the mustard-coated woman’s shoulder. The elderly attendant turns a small locked box toward her. She raises the brass key on its blue thread. End just before the key touches the lock.

Build a short drama one shot at a time

  1. Open with something the viewer can see. Give the lead a distinctive detail, such as a red sleeve or a silver ring, and repeat it in each prompt.
  2. Select a storyboard prompt or describe one action and one camera move. The form starts at 9:16. Check the duration, model and credit quote before generating each shot.
  3. Download the clips and arrange them in your editor. Cut on the reaction, add captions and end the episode just before the character answers.

Making duanju with AI

What does duanju mean?

Duanju is the pinyin for 短剧, meaning short drama. A vertical microdrama follows characters across short episodes, often ending on a reveal or an unresolved choice.

Can I create an entire episode here?

Each generation produces one clip. Make the opening, reaction and reveal as separate shots, then join them in a video editor to build your episode.

Which settings suit a vertical microdrama?

Start with 9:16 and five seconds for a single action. Use a close-up for a reveal, leave space for captions, and keep character and wardrobe details consistent across prompts.

How many credits does a scene use?

At five seconds and 480p, the signed-in Grok text model uses 10 Animate credits and Seedance 2.5 Flash uses 93. Check the quote again after changing duration or resolution.