Video prompt preparation

Turn your short video script into assets and timed video prompts

A prompt needs more than “make a warm cinematic video.” Specify where the subject stands, what they do, which image supplies each reference and how the camera follows the action. DramaForge prepares this package from your latest saved voiceover and shot list.

DramaForge · Updated

Choose a clip length, not a fixed cutting rhythm

A 60-second draft can become four 15-second plans or two 30-second plans. Each clip can contain several cuts, commonly four to six for 15 seconds or eight to twelve for 30, but dialogue may need longer shots. A 30-second plan does not mean every model can generate 30 seconds at once; split it to fit the selected tool.

Prepare reusable references first

Character prompts describe appearance, hair, age and clothing. Location prompts describe layout, time and lighting. Prop prompts describe shape, material and scale. Reuse one identity across clips, and assign consecutive local @image numbers with explicit reference roles rather than listing names without their purpose.

Include the words and sounds in the copied prompt

Use six sections: positions, reference assignments, timed cuts, lighting, sound design and continuity constraints. Put exact narration or dialogue inside the prompt, distinguish voiceover from lip-sync, and describe effects and music. Carry the final posture and prop state into the next clip. Regenerate when the saved script changes.

  • Keep the original narration
  • Start each local timeline at zero
  • Separate image prompts from video prompts
  • Check the target tool’s duration and audio support

Generation Sample

Original Illustrative Example

Write your own script

Input Material

Original example: A renter photographs the fridge and writes a moving checklist.

Structure Output

Character: renter in everyday clothing.
Location: kitchen and adjacent table.
Props: phone, paper, pen.
Clip: 15 seconds; voiceover, not lip-sync.

Prompt / Script Sample

Positions: Daytime kitchen. The renter stands facing the fridge, phone in right hand.
References: @image1 = renter; @image2 = kitchen; @image3 = phone.
Shot 1: 0-3s, medium shot, @image1 opens the fridge. Voiceover {Moving tomorrow? Check your fridge.}
Shot 2: 3-7s, hand close-up, @image3 photographs each compartment. <Soft shutter click>. Voiceover {Photograph what is still inside.}
Shot 3: 7-11s, overhead shot of paper and pen. Voiceover {Make a list before deciding what to take.}
Shot 4: 11-15s, medium shot, the renter places the pen beside the paper. Voiceover {Leave fewer decisions for moving day.}
Lighting: Soft natural daylight. Sound: Calm English voiceover, quiet room tone, no music.
Constraints: Consistent face and clothing, no subtitles, logos or watermarks; keep the paper visible in the final frame.

Workflow

Recommended Workflow

01

Save the draft

Confirm the latest narration and shot list.

02

Select clip duration

Choose 15 or 30 seconds and check tool limits.

03

Prepare assets

Review character, location and prop image prompts.

04

Review each clip

Check speech, timing, camera action and sound.

05

Copy or edit

Copy complete prompts, save manual edits or regenerate.

FAQ

Questions Creators Ask

Questions about sources, editing, credits and video prompts.

Is one clip the same as one shot?

No. A clip is a generation or assembly unit and may contain multiple shots. Cuts should follow the action and narration rather than a rigid two-second rule.

Will a 30-second prompt work in every model?

Not necessarily. Check the tool’s supported duration, references and audio features. Split the plan when needed and adapt syntax to the target model.

Does it use the latest edited script?

Generation uses the latest saved draft at launch. Editing the script afterwards flags the old plan as outdated.