ComfyUI-ProductPromptGenerator

ComfyUI-ProductPromptGenerator

A custom node for ComfyUI designed to generate structured prompts for product display images based on analysis results from Vision Language Models (like Qwen2.5-VL).

This node is particularly useful for e-commerce workflows where you want to automatically generate image-to-image prompts based on product characteristics (wearable, handheld, etc.).

Features

• Automated Prompt Logic: Automatically determines the best action verb ("wear", "hold", or "use") based on product attributes.
• JSON Parsing: Robustly parses JSON outputs from LLM/VLM nodes (handles markdown code blocks).
• Context Aware: Integrates product type, background, and interaction details into a cohesive natural language prompt.

Installation

1. Navigate to your ComfyUI custom nodes directory:

   cd ComfyUI/custom_nodes/
   

2. Clone this repository:

   git clone https://github.com/yourusername/ComfyUI-ProductPromptGenerator.git
   

3. Restart ComfyUI.

Usage

The node ProductPromptGenerator(Simple) accepts three string inputs, typically outputs from a VLM (Vision Language Model) that analyzes a product image.

Inputs

1. qwen_output_1 (String): Analysis of product type and background.

   • Expected JSON format:

     {
         "商品类型": "Ring",
         "是否支持手持": "No",
         "推荐背景": "Jewelry Store"
     }
     

2. qwen_output_2 (String): Analysis of interaction.

   • Expected JSON format:

     {
         "交互方式": "wearing on a finger"
     }
     

3. qwen_output_3 (String): Analysis of wearability.

   • Expected JSON format:

     {
         "是否支持佩戴": "Yes"
     }
     

Output

• prompt (String): A fully constructed prompt ready for CLIP Text Encode or other conditioning nodes.
  • Example Output:

    "请让图中模特佩戴图中戒指，背景为珠宝店，要求保持图中戒指与模特相对比例、大小、位置不变，画面聚焦模特与商品，人脸细节必须完整，戒指正确佩戴在模特的一根手指上。"

Example Logic

• If 是否支持佩戴 is "Yes" -> Verb: "佩戴" (Wear)
• If 是否支持手持 is "Yes" -> Verb: "手持" (Hold)
• Default -> Verb: "使用" (Use)