JSON gives you a structured description of visible style. The original prompt and hidden camera or generation settings cannot be recovered from the image.
1. Describe the reference image
Upload the picture whose style you like to ChatGPT, then paste this prompt.
Analyze this reference image and describe its visual style as reusable JSON. Include the medium, color palette, lighting, composition, background, textures, mood, and typography if present. Separate the subject from the style so I can apply the style to another picture. Describe only visible characteristics; mark uncertain details instead of inventing camera or generation settings. Return valid JSON.
2. Apply that style to your photo
Attach the photo you want to edit in ChatGPT. Replace the brackets with the JSON, the changes you want, and the details to preserve.
Edit the target photo I attached using the visible style described in the JSON below. Apply the lighting, color palette, textures, and mood where they fit this photo. Keep the person recognizable and preserve their facial features, expression, and body proportions. Preserve any details I list below. Use the subject description from the reference only as context.
Style JSON:
[Paste the JSON here]
Changes I want:
[Describe the edit]
Details to preserve:
[List details such as clothing, background, framing, or natural skin texture]