SuperAPI AI Directory

storyboard-breaker

Design & creation

View repository

Updated 10-15 17:33

About this skill

Storyboard Breakdown Guide

Core Definition: Storyboard Segment

One storyboard = one storyboard segment(segment) = one video generation task.

  • Each segment lasts 8-15 seconds, internally carrying 2-4 sub-shots
  • Between sub-shotscuts are allowed: change shot size, change angle, change subject, connect with hard cuts
  • Between sub-shotsno crossing scenes: a segment only takes place within one scene (scene_id is bound at the segment level)
  • Each sub-shot is 2-6 seconds, focusing on one visual unit (one action, one reaction, one close-up)

Splitting Process (Four Steps)

  1. Call read_storyboard_context to read the script, characters, scenes, props, and existing storyboard summaries
  2. Beat recognition: first identify the script's narrative beats—markers in the script such as [opening] [trigger] [climax] [ending], or narrative turning points (location shift, rule revelation, emotional outburst, reversal).Beat boundaries force segmentation, sub-shots within the same beat are prioritized into the same segment, do not split one causal chain (setup-occurrence-reaction) across different segments
  3. Total amount anchoring: target total duration = script word count ÷ 500 words/minute; segment count ≈ target total duration ÷ 12 seconds, allowing ±20% fluctuation. Do not significantly exceed or fall short
  4. Split sub-shots within a segment: split sub-shots by action switching points, perspective switching points, and subject switching points, complete all fields for each segment, then call save_storyboards to save all at once

Rhythm-layered duration

Determine duration by segment function, do not use a one-size-fits-all approach:

Segment type Duration Description
Transition segment 8-10 seconds Traveling, empty shots, environment establishment, transitions
Narrative segment 10-15 seconds Regular plot progression, dialogue
Climax segment 12-15 seconds Close-ups, rule revelation, emotional outburst, reversal; sub-shot rhythm slows down, a single sub-shot can linger for 4-6 seconds

Minimum dialogue duration (hard rule)

Segment duration ≥ total word count of dialogue and narration within the segment (the part written in description) ÷ 4.5 words/second + 2 seconds performance margin

Dialogue that does not fit must be split into the next segment; it is not allowed to cram dialogue that cannot be performed into one segment.

Shot elements

  1. Shot title: 3-5 words summarizing the segment's core content (e.g., "waking from a nightmare")
  2. Time: specific hour and minute + lighting description
  3. Location: complete scene description + spatial layout + environmental details
  4. Shot size: dominant shot size within the segment; for multi-shot-size segments write the combination, e.g., "medium shot + close-up"
  5. Angle: eye-level/low angle/high angle/side/back
  6. Camera movement: static/push in/pull out/pan/track/dolly (different sub-shots within a segment may differ)
  7. Visual description description: By 【镜头1】…【镜头2】… describe shot by shot what the audience actually sees and hears—the visuals (who + specific actions + physical details + expressions) come first; when the sub-shot has dialogue, write it as "Character Name says: "Dialogue"" in the corresponding 【镜头N】 , and write narration as "Narration: content"
  8. Visual Result result: the immediate consequences at the end of the segment + visual details
  9. Atmosphere atmosphere: lighting + color tone + sound + overall atmosphere
  10. Duration duration: the total duration of the segment is 8-15 seconds, and it meets the lower limit for dialogue duration
  11. Scene Association: if it can match an existing scene, you must fill in scene_id
  12. Character Association: fill in character_ids, binding 0 to multiple characters involved in the current segment
  13. Prop Association: fill in prop_ids, binding the key props appearing in the current segment (0 to multiple)

Scene Association Rules

  • Prioritize using read_storyboard_context returned by scenes
  • location + time When an exact match can be made, you must fill in the correct scene_id
  • Do not fabricate nonexistent scene IDs
  • If the script content clearly falls within an existing scene, do not repeatedly create new scene descriptions

Character Binding Rules

  • character_ids must be selected from read_storyboard_context the returned character list
  • A segment can have no characters, or it can bind multiple characters
  • As long as there are characters who clearly appear, are seen, perform actions, or speak in the segment, they should all be bound
  • Pure environment segments, empty shots, and object close-ups can pass an empty array

Prop Binding Rules

  • prop_ids must be selected from read_storyboard_context the returned prop list (props)
  • When a prop is used by a character, handed over, shown in close-up, or is clearly visible in the frame and meaningful to the narrative, it must be bound to that segment
  • Prop close-up segments (with no characters) should also bind props,character_ids can be empty
  • Do not bind background items or scene furnishings unrelated to the plot; for segments with no props appearing, pass an empty array
  • Bound props will serve as reference images for video generation (single-item images on a white background), ensuring prop appearance consistency across segments

Quality Requirements

  • description should be suitable for human reading, describing shot by shot in detail what the audience actually sees and hears; dialogue/narration should be written directly in the corresponding 【镜头N】 within
  • image_prompt should highlight the single-frame composition, character appearance, environment, and lighting (corresponding to the first sub-shot of the segment)
  • bgm_prompt and sound_effect can use concise phrases, but should not be so vague that they are only "tense" or "sad"
  • If adjustments are needed, call update_storyboard to modify the specific segment