1. Workflow Overview

This workflow demonstrates how to call OpenAI's GPT-Image-1 model through ComfyUI, featuring:

• Professional style transformation while preserving facial features

• Automatic layout generation (9:16 vertical format)

• Detailed prompt control for fashion elements

2. Node Breakdown

Key Nodes:

1. OpenAIGPTImage1

   • Core node for GPT-Image-1 API calls

   • Critical parameters:

     • Prompt: Detailed occupational style description

     • Resolution: 1024x1536

     • Quality: high

2. LoadImage

   • Input portrait (800x1449px)

3. PreviewImage

   • Real-time result preview

Note:

• Requires local ComfyUI latest version

• OpenAI API key needed

3. Workflow Structure

Processing Stages:

1. Input

   • Load source portrait

2. Generation

   • GPT-Image-1 transforms outfit based on:

     "Generate 00TD occupational styles...First profession: Fashion designer"

3. Output

   • Preview generated fashion set

4. I/O Specification

Inputs:

• Source image: Vertical portrait

• Prompt: Detailed clothing requirements

• Resolution: 1024x1536

Outputs:

• Generated image showing:

  • Left: Full-body styling

  • Right: Item breakdown

  • Top: "00TD" text mark

5. Critical Notes

1. Requirements

   • Latest ComfyUI mandatory

   • Valid OpenAI API key

2. Tips

   • Change profession in prompt:

     text

     复制

     下载

     "First profession: Financial analyst"

   • Adjust quality parameter as needed

3. Billing

   • GPT-Image-1 calls incur OpenAI API costs

   • Test prompts in Playground first