Sv2

Image Generation Prompt Template

Product: {product_description}

SCENE: {scene_setting}
POSE: {scene_pose}

CRITICAL REQUIREMENTS:
- Western/European model, standing, full body visible
- Product must be clearly visible and remain the focal point
- Exact product fidelity: same color, pattern, fabric texture, design details
- High-end lifestyle fashion editorial style
- Natural, candid feel — not stiff or overly posed

IMPORTANT: This image is part of a 2-image set for video generation. The model's face, hair, and body type must be consistent and recognizable across images.

Cross-Image Consistency Rules

Since both images are used as video keyframes (first frame / last frame), the following must remain identical across Scene A and Scene B:

1. Model identity: Same face, hair color, hairstyle, body type, skin tone
2. Product appearance: Identical garment — color, pattern, fabric texture, silhouette, construction details, fit
3. Photography quality: Same resolution, similar framing (full body), professional editorial grade

Only the scene/environment, lighting, and pose should differ between the two images.

Material-Specific Notes

• Denim/Cotton: Emphasize texture and structure; works well in both urban and natural settings
• Silk/Satin: Capture fabric flow and sheen; golden hour lighting enhances luster
• Knit/Wool: Highlight texture and warmth; overcast light shows knit detail
• Sheer/Lightweight: Show movement and drape; walking pose in Scene B is ideal
• Leather/Structured: Emphasize sharp lines; urban backdrop in Scene A complements best

Image Analysis Prompt

You are an expert fashion videographer analyzing e-commerce model images.
Analyze the provided image and extract structured JSON:
{
  "model": { "pose": "", "expression": "", "position": "" },
  "clothing": {
    "type": "", "fit": "", "fabric": "", "keyFeatures": [],
    "highlightAreas": ["2-3 most visually striking areas"]
  },
  "scene": { "environment": "", "background": "", "props": "" },
  "lighting": { "type": "", "direction": "", "mood": "" },
  "camera": { "angle": "", "framing": "" },
  "videoActions": {
    "shot1_action": "recommended action for establishing shot",
    "shot2_details": ["2-3 detail areas for close-up"],
    "shot2_handActions": ["2-3 hand interactions"],
    "shot3_action": "closing pose/action"
  }
}