SuperAPI AI Directory
Partner中文AI newsFavorites
SkillsMiniMax-H3-Storyboard-Prompt-Generator

MiniMax-H3-Storyboard-Prompt-Generator

Agents & models

View repository

Updated 10-16 08:42

About this skill

MiniMax H3 Storyboard Prompt Generation Skill

Skill Objective

Based on user needs, generate a structured storyboard prompt that conforms to the MiniMax H3 official specification, ensuring continuity, physical plausibility, and media specification compliance during video generation.

H3 Generation Mode Overview

Model Variants Input Modes Specifications
H3-Base-FL2VA(first-last frame mode) 0~2 images · 0 images: text-to-video· 1 image: first-frame or last-frame constrained generation· 2 images: first-and-last-frame constrained generation
H3-Base-Ref2VA(full reference mode) images ≤9, videos ≤3 (total duration ≤15s), audio ≤3 (uploaded together with images/videos) total files ≤12; each video clip 215s; each audio segment 215s; output duration 4~15s, 24fps, stereo 32kHz

Output specifications are fixed: duration 4–15 seconds, aspect ratio 21:9 / 16:9 / 4:3 / 1:1 / 3:4 / 9:16, resolution defaults to 768px on the short side (2K can be enabled), frame rate 24fps.

Prompt Construction

Preserve the user's intent. When the user's description is not detailed enough, expand it into a production-oriented prompt in the following order:

  1. Output target: duration, aspect ratio, purpose, and whether the segment should feel like a continuous shot or an edited sequence.
  2. Subjects and assets: specify the characters, products, and locations, whether reference images/videos/audio are used, what each reference image/video/audio contains (have the user refer to reference content in the form of image 1, image 2, video 1, video 2…, with a brief description), and how they should be used.
  3. Timeline: describe actions in chronological order, with clear transitions.
  4. Scene: environment, time of day, lighting, weather, and important background details.
  5. Camera: shot size, angle, movement, focus behavior, and editing approach.
  6. Visual style: color, texture, mood, pacing.
  7. Sound: dialogue, ambient sound, music, sound effects, and how they sync with the reference audio.
  8. Constraints: elements to preserve and flaws to avoid, such as identity drift, flickering, text, logos, or abrupt cuts.

When the user has already provided a complete structured storyboard script, keep it as is. Do not compress, translate, or replace it with a generic prompt. Only check the prompt length, timestamp coverage, reference numbering, media compatibility, and direct contradictions before submission.

Operational heuristics for better results:

  • Ensure the action is physically feasible within the chosen 4–15 second duration.
  • Prioritize one clear action beat per shot, rather than cramming unrelated events into the same block.
  • Make timestamp ranges continuous, with no gaps or overlaps, and the final end time equal to the requested duration.
  • Use a two-level timeline: each reference image has one main shot range, then split micro-ranges within that shot for preparation, execution, stabilization, and the final hold.
  • Use 0.5-second precision by default. Use finer timing only when precise physical transitions within a short time are needed.
  • Each shot must declare: the exact range and duration, reference image number, initial state, shot/camera, micro-timeline, and the locked end state.
  • The locked end state of shot N must be the initial state of shot N+1. Check character pose, gaze, which hand is holding, prop ownership/position, prop connection relationships, and camera side.
  • Explain what must remain consistent and what may change. For example: preserve the face, clothing, and product shape; change the pose, camera angle, and background motion.
  • When important, explicitly lock the persistent spatial relations: left/right seating, foreground/background, screen direction, and which hand is holding the object.
  • For handoffs and precise object motion, describe the state transitions in causal order: initial owner and position, release condition, transfer, new owner, final state. Do not compress these into one vague action.
  • Separate hard continuity constraints from aesthetic preferences and negative constraints. When breaking a key invariant would destroy the scene, repeat that invariant in the relevant timeline block.
  • Use explicit camera language, such as static shot, close-up, tracking shot, push-in, pan, tilt, handheld, or aerial.
  • Specify audio timing only when audio is provided. State whether motion, cuts, lip-sync, or sound effects should follow the beat or the spoken words.
  • Do not arbitrarily add brands, dialogue, text overlays, or unsafe content that the user did not request.

Timeline planning process:

  1. Read the total duration D and the number of ordered storyboard shots N。
  2. Preserve the timestamps provided by the user. Otherwise, allocate D / N seconds to each storyboard shot; for example, 15 seconds, 6 images, then initially six 2.5-second shots.
  3. Reallocate time based on action complexity while keeping all ranges continuous and totaling exactly D. Allocate more time to precise handoffs or multi-stage physical actions; allocate less time to static establishing shots or reaction shots.
  4. Split each shot into 3–5 readable micro-beats: establish/hold, preparation, core action, stabilize/recover, and an optional final hold.
  5. Maintain a state ledger at each boundary: character position and pose, gaze, left/right hand state, prop owner and position, physical connection relationships, and camera side.
  6. Verify that each reference is used exactly once in the input order (unless the user explicitly requests reuse), and that the locked end state of shot N exactly matches the initial state of shot N+1.
  7. If the requested action cannot fit without being rushed or vague, simplify the action plan or inform the user before initiating a paid request. Do not silently compress causal steps.

For two or more ordered storyboard images, use the following structure. Do not flatten it into a single paragraph.

Chinese structured storyboard template:

【输出规格】
生成一段{duration}秒、{ratio}的{用途}视频。使用{N}张分镜参考图，严格按输入顺序作为连续镜头参考。按时间轴依次生成，不跳镜，不改变已锁定的人物、服装、位置和核心道具。

【整体风格】
{真人/动画/产品等媒介}，{电影或视觉风格}，{光线、色彩、纹理、景深、节奏和运镜基调}。场景为{环境、时间、天气和背景运动}。声音为{对白、环境声、音乐或静音要求}。

【人物与空间连续性】
人物1：{外观、服装、气质}，始终位于{固定位置}。
人物2：{外观、服装、气质}，始终位于{固定位置}。
全程保持{脸、发型、服装、座位、朝向、比例和场景结构等硬约束}。

【道具状态连续性】
全片只有{道具数量和名称}。初始状态：{所有者、所在手、位置、连接关系}。过程中必须保持{形状、数量、连接、可见性和物理行为}，不得凭空出现、消失、复制、跳位或改变所有者。

【两级时间轴与参考图映射】
镜头1｜0:00-{T1}｜时长{D1}秒｜对应第1张图
镜头与运镜：{景别、角度、镜头运动和对焦}。
起始状态：{人物姿势、视线、双手状态、道具所有者/位置/连接关系}。
镜头内微时间轴：
- 0:00-{t1a}：建立画面，{可见状态和轻微环境运动}。
- {t1a}-{t1b}：预备动作，{视线、重心、手部或道具开始变化}。
- {t1b}-{t1c}：核心动作，{只执行本镜头的主要动作}。
- {t1c}-{T1}：动作完成后稳定并保持，禁止提前执行下一镜头动作。
结束状态锁定：{逐项写明人物、手、道具和连接关系，供镜头2直接继承}。

镜头2｜{T1}-{T2}｜时长{D2}秒｜对应第2张图
镜头与运镜：{景别、角度、镜头运动和对焦}。
起始状态：必须与镜头1的结束状态完全一致。
镜头内微时间轴：
- {T1}-{t2a}：保持上一镜头状态，给动作留出可见停顿。
- {t2a}-{t2b}：{预备动作}。
- {t2b}-{t2c}：{核心动作}。
- {t2c}-{T2}：{完成、回收或保持动作}。
结束状态锁定：{明确状态}。

{继续为每张参考图写一个完整镜头块。主时间轴和每个镜头内的微时间轴都必须连续无空档，最后一个时间点必须等于duration。}

【严格动作顺序】
{动作A} → {动作B} → {动作C} → {动作D}。不得合并、倒序或省略。每个镜头只完成一个明确动作；涉及交接时，先写谁保持不动，再写谁接触并拿稳，最后写原持有者何时松手。

【负面约束】
不要{不需要的视觉类型、剪辑风格和表演方式}。全程无{字幕、水印、多余人物或多余道具}，避免{身份漂移、变装、换位、闪烁、畸形手指、多余肢体、物体跳变、连接关系错误和突兀切镜}。

English structured storyboard template:

[OUTPUT SPECIFICATION]
Create a {duration}-second, {ratio} {use case} video using {N} storyboard reference images in their exact input order as consecutive shot references. Follow the timeline without skipping shots or changing locked characters, wardrobe, positions, or key props.

[OVERALL LOOK]
{Live action/animation/product medium}; {cinematic or visual style}; {lighting, color, texture, depth of field, pacing, and camera baseline}. Scene: {environment, time, weather, and background motion}. Sound: {dialogue, ambience, music, reference-audio synchronization, or silence}.

[CHARACTER AND SPATIAL CONTINUITY]
Character 1: {appearance, wardrobe, demeanor}; always at {fixed position}.
Character 2: {appearance, wardrobe, demeanor}; always at {fixed position}.
Preserve {faces, hair, wardrobe, seats, screen direction, proportions, and scene layout} throughout.

[PROP STATE CONTINUITY]
The entire video contains exactly {prop count and names}. Initial state: {owner, hand, location, and connection}. Preserve {shape, count, connection, visibility, and physical behavior}; never duplicate, remove, teleport, disconnect, or silently change ownership.

[TWO-LEVEL TIMELINE AND REFERENCE MAPPING]
Shot 1 | 0:00-{T1} | Duration {D1}s | Reference image 1
Shot and camera: {framing, angle, movement, and focus}.
Initial state: {poses, gazes, both-hand states, prop owner/location/connections}.
In-shot micro-timeline:
- 0:00-{t1a}: Establish the frame with {visible state and subtle environment motion}.
- {t1a}-{t1b}: Preparation: {gaze, weight, hand, or prop begins to change}.
- {t1b}-{t1c}: Core action: {perform only this shot's main action}.
- {t1c}-{T1}: Settle and hold the completed action; do not begin the next shot's action early.
Locked end state: {explicit character, hand, prop, and connection state inherited by shot 2}.

Shot 2 | {T1}-{T2} | Duration {D2}s | Reference image 2
Shot and camera: {framing, angle, movement, and focus}.
Initial state: exactly match shot 1's locked end state.
In-shot micro-timeline:
- {T1}-{t2a}: Hold the inherited state long enough to remain readable.
- {t2a}-{t2b}: {preparation}.
- {t2b}-{t2c}: {core action}.
- {t2c}-{T2}: {completion, recovery, or hold}.
Locked end state: {explicit state}.

{Continue with one complete shot block per reference image. Both the master timeline and every in-shot micro-timeline must be contiguous; the final timestamp must equal duration.}

[STRICT ACTION ORDER]
{Action A} -> {Action B} -> {Action C} -> {Action D}. Do not merge, reorder, or omit steps. Each shot completes one clear action. For a handoff, state who remains still, who establishes and secures contact, and only then when the original holder releases.

[NEGATIVE CONSTRAINTS]
No {unwanted visual genre, editing style, or performance}. No {subtitles, watermarks, extra people, or extra props}. Avoid {identity drift, wardrobe changes, position swaps, flicker, malformed fingers, extra limbs, prop teleportation, broken connections, and abrupt cuts}.

For a single text-to-video request or a single loose reference image, one concise paragraph is sufficient. When continuity, multiple ordered images, or precise physical interaction is critical, use the full storyboard structure.

Detailed timing example: headphone removal and handoff

Chinese:

镜头2｜0:02.5-0:05.0｜时长2.5秒｜对应第2张图
起始状态：女生低头看手机，两只耳塞都在女生耳中；手机在女生手中；男生双手仍抱着书包。
0:02.5-0:03.0：女生把手机稍微放低并短暂看向男生，其他状态不变。
0:03.0-0:04.1：女生用靠近男生的手，以拇指和食指捏住耳机柄，缓慢从自己耳中取下一个耳塞。
0:04.1-0:05.0：女生把耳塞停在两人之间、胸口高度；男生尚未伸手；耳塞与男生脸和耳朵之间保留明显空隙。
结束状态锁定：女生仍捏着耳机柄，另一只耳塞仍在女生耳中，耳机线仍连接女生手机；男生双手仍在书包上。

镜头3｜0:05.0-0:07.5｜时长2.5秒｜对应第3张图
起始状态：完全继承镜头2的结束状态。
0:05.0-0:05.5：女生保持手和耳塞不动；男生短暂停顿。
0:05.5-0:06.4：男生抬起靠近女生的手，拇指和食指接近耳机柄，女生仍不松手。
0:06.4-0:07.0：男生夹住并拿稳耳机柄；只有确认拿稳后，女生才松手。
0:07.0-0:07.5：女生把手收回；男生持有耳塞，但尚未把它移向耳朵。
结束状态锁定：耳塞所有者已变为男生，位于两人之间；女生手已退出交接区域；耳机线仍从女生手机自然连接到耳塞。

English:

Shot 2 | 0:02.5-0:05.0 | Duration 2.5s | Reference image 2
Initial state: the girl looks at her phone with both earbuds in her ears; she holds the phone; the boy still holds his bag with both hands.
0:02.5-0:03.0: The girl lowers the phone slightly and briefly looks at the boy; all other states remain unchanged.
0:03.0-0:04.1: With the hand closer to the boy, she pinches one earbud stem between thumb and index finger and slowly removes it from her ear.
0:04.1-0:05.0: She holds the earbud at chest height between them; the boy has not reached for it; a visible gap remains between the earbud and his face and ear.
Locked end state: the girl still holds the earbud stem; the other earbud remains in her ear; the cable remains connected to her phone; both of the boy's hands remain on his bag.

Shot 3 | 0:05.0-0:07.5 | Duration 2.5s | Reference image 3
Initial state: exactly inherit shot 2's locked end state.
0:05.0-0:05.5: The girl keeps her hand and the earbud still; the boy pauses.
0:05.5-0:06.4: The boy raises the hand closer to the girl and approaches the earbud stem with thumb and index finger; she does not release it.
0:06.4-0:07.0: The boy grips and secures the earbud stem; only after his grip is stable does the girl release it.
0:07.0-0:07.5: The girl withdraws her hand; the boy holds the earbud but does not move it toward his ear yet.
Locked end state: ownership has transferred to the boy; the earbud remains between them; the girl's hand has left the handoff area; the cable still connects naturally to the girl's phone.