SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillsbyted-ind-ecom-product-video-prompt

byted-ind-ecom-product-video-prompt

Design & creation

View repository

Updated 10-15 17:30

About this skill

Product Video Prompt Generator

This skill uses Seedance 2.0 "three-layer skeleton" methodology to convert static product images into professional video generation prompts. It provides structured templates, category-specific examples, and quality standards for creating compelling e-commerce product videos.

When to Use This Skill

  • Convert single or multiple product images into video prompts
  • Create professional product showcase video scripts
  • Generate structured prompts for Seedance 2.0 video generation
  • Need category-specific templates (footwear and apparel, bags, beauty, 3C accessories)
  • Want to follow the industry-standard "three-layer skeleton" methodology

Methodology Overview

The "three-layer skeleton" methodology includes:

  1. Global Settings (Layer 1): Define the product, characters, scene, composition, lighting, and style
  2. Time-Segment Storyboard (Layer 2): Use the "five-segment storyboard framework" to design the narrative rhythm (0-3s, 3-6s, 6-9s, 9-12s, 12-15s)
  3. Constraints and Output Specifications (Layer 3): Set consistency rules, quality standards, and parameters

Quick Start: 30-Second Recipe

  1. Choose a template: Single-image template or multi-image template
  2. Fill in the core slots: Product name + reference images
  3. Define key elements: Characters, scene, lighting
  4. Generate: Use the template default settings for the first version
  5. Iterate: Adjust based on the results

Templates

Template 1: Generate Video from a Single Image

Suitable for scenarios where there is only one core product image and AI is needed to creatively extend around it.

## 全局设定
- **商品**: [你的商品名称] (参考[图1])
- **人物**: [模特风格、气质、着装描述，或"无人物"]
- **场景**: [具体场景描述，如"简约明亮的原木风室内工作室"]
- **构图**: [构图风格，如"中心构图，画面极简平衡"]
- **光线**: [光线描述，如"清晨自然侧向柔光"]
- **风格**: 写实风格，真实拍摄，电影级质感

## 时间片分镜脚本
1. **0s-3s (引出主体)**: [远景/全景]展示[模特状态或商品陈列]，镜头[推/拉/移]至[中景/近景]，呈现[商品名称]的整体[廓形/外观]。
2. **3s-6s (核心质感)**: [中景/近景]拍摄[模特动作或商品细节]，展示[商品材质/设计特点]，[固定机位/慢速推近]。
3. **6s-9s (场景交互)**: [中景/全景]捕捉[商品与环境元素（如光影、微风、道具）的互动]，展现其[动态美/氛围感]。
4. **9s-12s (多角度展示)**: [中景/近景]采用[环绕/旋转/弧线]运镜，多角度呈现[商品名称]的[立体感/侧面细节]。
5. **12s-15s (品牌收尾)**: [全景/中景]定格[模特或商品]在[场景描述]下的最终画面，营造[品牌氛围，如"高端极简"或"活力时尚"]氛围。

## 约束与输出规格
- **一致性**: [商品名称]必须与参考图[图1]严格一致，无变形、无穿模。
- **人物稳定性 (如需要)**: [人物描述]动作真实流畅，面部五官稳定清晰，无异常姿态，手部无畸变。
- **美学要求**: 8K超清画质，光影自然通透，色彩真实，背景适当虚化。
- **Seedance 2.0 参数**: 时长[15s], 画幅[16:9], 风格强度[中], 节奏[舒缓]。

Template 2: Generate Video from Multiple Images

Suitable for scenarios where product images from multiple angles are provided and AI is needed to strictly maintain consistency of features across all angles.

## 全局设定
- **商品**: [你的商品名称] (参考[图1], [图2], [图3]...)
- **人物**: [模特风格、气质、着装描述，或"无人物"]
- **场景**: [具体场景描述]
- **构图**: [构图风格]
- **光线**: [光线描述]
- **风格**: 写实风格，真实拍摄，商业大片质感

## 时间片分镜脚本
1. **0s-3s (引出主体)**: [全景/远景]展示[商品名称](参考[图1]的整体外观)在[场景]中，镜头[拉远/推近]，展示商品与空间的[开阔感/和谐感]。
2. **3s-6s (核心质感)**: [中景/近景]侧面拍摄[模特穿着/使用着商品名称]，展示其[关键特点如"厚底廓形"或"皮革纹理"]，(参考[图2])。
3. **6s-9s (场景交互)**: 中景捕捉[模特/商品]在[场景元素如"窗边"]位置，[光线/环境元素]在商品表面产生[效果]，(参考[图3])，映射出[材质/设计]的层次变化。
4. **9s-12s (多角度展示)**: [中景/全景]采用[环绕/摇臂]等运镜方式，全面展示[商品名称]的(参考[图4])[空间立体轮廓/多维细节]。
5. **12s-15s (品牌收尾)**: [全景]镜头下，[商品名称](参考[图5])定格在[光影环境]中，[背景虚化/焦点]处理，营造出[品牌调性]氛围。

## 约束与输出规格
- **一致性**: [商品名称]必须在所有参考角度([图1]-[图5])之间严格保持特征一致，细节、颜色、Logo完全匹配，无穿模。
- **人物稳定性 (如需要)**: [人物描述]动作连贯自然，面部五官稳定清晰，表情一致，无手部畸变。
- **美学要求**: 超清8K画质，光线通透，色彩精准。
- **Seedance 2.0 参数**: 时长[15s], 画幅[9:16], 风格强度[高], 节奏[动感]。

Category-Specific Examples

Category 1: Footwear and Apparel (sneakers, with characters)

## 全局设定
- **商品**: A品牌运动休闲鞋 (参考[图1], [图2], [图3])
- **人物**: 一位阳光活力的时尚运动风男模
- **场景**: 极简主义都市现代工业风展厅，带有几何展示台和大落地窗
- **构图**: 中心构图配合稳定的平衡构图
- **光线**: 通透的自然柔光与细腻边缘光
- **风格**: 写实风格，真实拍摄，商业广告级质感

## 时间片分镜脚本
1. **0s-3s (引出主体)**: 全景展示白色运动鞋(参考[图1])置于极简几何展示台上，镜头平滑拉远，展示产品与空间的开阔感。
2. **3s-6s (核心质感)**: 中景侧面拍摄男模稳健迈步穿着鞋子(参考[图2])，镜头平移跟随，强调厚底廓形与落地时的稳重支撑感。
3. **6s-9s (场景交互)**: 中景捕捉男模站立于窗边，自然午后光影在鞋面缓缓流动，映射出鞋身材质(参考[图3])的层次变化。
4. **9s-12s (多角度展示)**: 运用中景全景环绕运镜，随着镜头平滑右移，全面展示运动鞋的空间立体轮廓与运动线条。
5. **12s-15s (品牌收尾)**: 全景镜头下，产品定格于优雅光影环境中，背景虚化处理，营造出高端极简时尚休闲氛围。

## 约束与输出规格
- **一致性**: 视频中的运动休闲鞋需要与输入参考图片严格保持一致。
- **人物稳定性**: 模特动作真实无异常，面部五官清晰，无手部畸变。
- **美学要求**: 超清8K画质，真实风格。
- **Seedance 2.0 参数**: 时长15s, 画幅16:9, 严格遵循分镜脚本描述。

Category 2: 3C Accessories (phone case, conceptual style)

## 全局设定
- **商品**: X品牌未来科技感手机壳 (参考[图1])
- **人物**: 无需展示全身人物，仅需干净优雅的手部特写
- **场景**: 赛博朋克风格的城市夜景，充满霓虹灯光和金属质感
- **构图**: 特写和中特写相结合，配合中心构图，强调产品细节
- **光线**: 高对比度的戏剧性光影效果，如锐利的边缘光、动态光束
- **风格**: 写实风格，真实拍摄，带有未来科技感

## 时间片分镜脚本
1. **0s-3s (引出主体)**: 特写在黑暗背景中一束光扫过手机壳的一角，缓缓揭示其金属光泽和精准纹理；配合慢速拉远镜头，突显未来科技质感并营造悬念。
2. **3s-6s (核心质感与贴合)**: 中特写展示手机安装过程，聚焦手机壳与手机边框的精准贴合；配合稳定推入镜头和轻微"咔嗒"声效，强化严谨工艺与高度适配性。
3. **6s-9s (场景交互与氛围)**: 中特写到近中景展示手持手机，在充满霓虹灯的未来城市夜景中滑动屏幕；配合轻柔跟随镜头和霓虹灯光变幻，流光在手机壳磨砂表面流动，强化未来科技氛围。
4. **9s-12s (多角度展示)**: 中景到全景运用"子弹时间"式快速环绕镜头，展示手机在空中旋转；配合定格+环绕机位切换，全面展示手机壳的立体保护结构与纤薄设计。
5. **12s-15s (品牌收尾与记忆点)**: 特写手机静静放置于金属控制台上，一束顶光精准照亮中心品牌Logo；配合慢速收光，背景逐渐暗下，仅剩高亮度Logo，营造神秘、专业的科技收尾画面。

## 约束与输出规格
- **一致性**: X品牌未来科技感手机壳必须与参考图[图1]严格匹配，无变形，无穿模。
- **人物稳定性 (如需要)**: 手部动作必须真实流畅，如出现面部则面部五官必须稳定清晰，无异常姿态，无手部畸变。
- **美学要求**: 超清8K画质，自然通透光影，真实色彩；背景适当虚化以突出主体；手机壳品牌Logo需在需要呈现时始终清晰可见。
- **Seedance 2.0 参数**: 时长 [15s], 画幅 [9:16], 风格强度 [中], 节奏 [舒缓]。
- **严格遵循分镜脚本描述**

Category 3: Bags (handbag, no characters)

## 全局设定
- **商品**: B品牌帆布托特包 (参考[图1], [图2])
- **人物**: 无人物
- **场景**: 极简主义现代暖阳露台，带有木质咖啡桌和绿植
- **构图**: 经典平衡中心构图
- **光线**: 自然明亮侧逆柔光
- **风格**: 写实风格，真实拍摄，生活杂志质感

## 时间片分镜脚本
1. **0s-3s (引出主体)**: 全景镜头下，多个托特包(参考[图1])错落有致地摆放在纯净石阶上，镜头平滑推入，展示产品整体廓形与场景和谐美感。
2. **3s-6s (核心质感)**: 中景，粉色格纹包(参考[图2])静静置于木质咖啡桌上，微风吹过，光影在帆布材质上流动，展示其挺括与柔软兼具的质感。
3. **6s-9s (场景交互)**: 特写镜头，焦点从背景虚化的绿植过渡到包的提手部分，展示其皮革缝线细节。
4. **9s-12s (多角度展示)**: 全景环绕镜头，从侧后方平滑弧线至正面，展示不同颜色包在空间的立体廓形与色彩层次。
5. **12s-15s (品牌收尾)**: 中景定格，多个包静静伫立于柔和的午后光影中，通过稳定镜头拉远，营造高端极简时尚品牌情调。

## 约束与输出规格
- **一致性**: 视频中的托特包需要与输入参考图片严格保持一致。
- **美学要求**: 超清8K画质，色调柔和，避免过度锐化。
- **Seedance 2.0 参数**: 时长15s, 画幅9:16, 严格遵循分镜脚本描述。

Category 4: Beauty (lipstick, hands on camera)

## 全局设定
- **商品**: C品牌丝绒口红 (参考[图1])
- **人物**: 无需展示全身人物，仅需干净优雅的女性手部特写
- **场景**: 柔光摄影棚，背景为纯净的裸粉色
- **构图**: 特写和中特写相结合，配合中心构图，强调产品细节
- **光线**: 干净柔和的摄影棚灯光，无明显阴影
- **风格**: 写实风格，真实拍摄，高端美妆广告级质感

## 时间片分镜脚本
1. **0s-3s (引出主体)**: 特写镜头，口红管身在纯净背景中缓缓旋出膏体，展示其完美的丝绒质地和精准切面，镜头缓缓推入。
2. **3s-6s (核心质感)**: 特写镜头，干净的手指轻轻抚过口红管身的哑光材质，感受其高端品质。
3. **6s-9s (场景交互)**: 特写镜头，口红膏体在手臂肌肤上滑过，留下一条色彩饱和、质地均匀的印记。
4. **9s-12s (多角度展示)**: 运用"子弹时间"式快速环绕镜头，展示口红在空中旋转，全面展示其设计细节和品牌Logo。
5. **12s-15s (品牌收尾)**: 口红静静放置于控制台上，一束顶光精准照亮中心品牌Logo，画面渐渐暗下，Logo保持高亮，营造神秘、专业的科技氛围。

## 约束与输出规格
- **一致性**: 视频中的口红需要与输入参考图片严格一致，特别是颜色和质地。
- **人物稳定性**: 手部动作必须流畅自然，无畸变，指甲干净。
- **美学要求**: 超清8K画质，确保口红色号准确还原，无色差。
- **Seedance 2.0 参数**: 时长15s, 画幅1:1, 严格遵循分镜脚本描述。

Pre-Generation Quality Checklist

Before generating, check your prompt against the following criteria:

  • Clarity: Instructions are simple and direct, with no ambiguous wording
  • Specificity: Provide enough detail (for example, "a cafe on a Paris street corner at 4 PM" vs "a nice scene")
  • Semantic Consistency: Global settings, storyboard script, and constraints are unified in style
  • Product Focus: All descriptions serve to "showcase the product"
  • Consistency Anchors: Use the strictest wording to require the product/characters to remain consistent with the reference images
  • Core Selling Points: The storyboard script clearly plans shots that showcase the key selling points
  • Style Authenticity: Explicitly specify "realistic, real-life shooting"
  • Format specifications: Confirm the aspect ratio and duration of the target platform

Risk control: common issues and countermeasures

Issue 1: Subject inconsistency

  • Symptom: The generated product differs from the reference image in color, shape, or logo
  • Solution: Strengthen the binding by using "the product in the video must be 100% strictly consistent with the reference image [Image 1] in all details such as form, color, logo, etc."

Issue 2: Character/geometry anomalies

  • Symptom: Extra fingers, unnatural poses, or the product bending unnaturally
  • Solution: Add negative constraints: "avoid any hand or facial distortion," "ensure the model's limb proportions are normal," "the product form must remain geometrically stable"

Issue 3: Over-stylization

  • Symptom: The image is too cartoonish or oil-painting-like, losing the commercial texture of "real shooting"
  • Solution: Fix the style as "realistic style, real shooting" and add "cinematic image quality" and "commercial advertising-grade texture"

Issue 4: Information occlusion

  • Symptom: The model's pose, hands, or objects in the scene block key parts of the product
  • Solution: Optimize the storyboard description: "the model naturally lets the bag hang at their side, avoiding hands blocking the logo on the bag"

Issue 5: Pacing and duration

  • Symptom: The overall pacing is too fast or too slow, or the duration does not meet expectations
  • Solution: Adjust the storyboard timestamps: change "0s-3s" to "0s-2s" to speed up the pacing, or change it to "0s-4s" to slow it down

Iteration strategy

Strategy 1: Fine-tune the storyboard description

If only one shot is unsatisfactory (wrong angle, too much movement), precisely modify that storyboard's核心内容or运镜方式. For example, change "the model turns around dramatically" to "the model turns slightly to the side"。

Strategy 2: Lock randomness, fix keyframes

If a certain frame (such as storyboard 3) is already perfect, but other storyboards need adjustment, you can try telling the AI "keep the visual style and composition of storyboard 3 unchanged", or provide a screenshot of that frame as a new reference image.

Strategy 3: Adjust style weight

If the overall atmosphere deviates from expectations (too "fake" or too bland), adjust场景、光线the strength of stylization descriptions such as. If it is too "fake," reduce modifiers and move toward "natural" and "minimalist."

Strategy 4: Make good use of "subtraction" and add negative constraints

If the same problem occurs repeatedly (such as hand distortion), do not repeatedly modify the positive description. Instead, directly add a "negative list" in the constraints section, clearly telling the AI "what not to do". Negative prompts are often more effective than positive guidance.

Usage workflow

  1. Analyze input: Determine whether the user has a single image or multiple images
  2. Choose template: Choose Template 1 (single image) or Template 2 (multiple images)
  3. Fill slots: Fill all [bracketed] placeholders with specific content
  4. Apply category guidance: If applicable, refer to category-specific examples
  5. Quality check: Validate according to the pre-generation checklist
  6. Generation and Iteration: Create a video, then iterate using optimization strategies

Output Format

This skill outputs complete, ready-to-use prompts that follow a three-layer skeleton structure:

  • Global settings that define all 6 elements
  • A 5-segment storyboard timeline with detailed shot descriptions
  • Constraints and technical specifications

This prompt can be used directly with Seedance 2.0 or similar video generation models.