Edit a photo by uploading one source image and describing the changes you want in a prompt. New AI generations replace the original; no pixel-level editing tools.
Edit a photo using GPT Image 2. Upload one source/reference photo and write a prompt that states what must remain and what should change. The output is a new AI generation based on your instructions. Use words in your prompt to request framing or crop adjustments.
Choose one clear photo to serve as the reference. The model uses this image to understand the subject, composition, and lighting.
Describe what should stay the same and what should change. For example: 'Keep the mug and its position exactly as shown. Replace the wooden table with a clean white studio backdrop.'
The tool produces a new image. Check that the preserved elements match the source and the requested changes are applied. Regenerate if needed.
A white ceramic mug on a wooden table. Change the background to a clean white studio backdrop. Keep the mug, its position, and the lighting on the mug exactly the same. The table should be replaced by the white backdrop.
The mug should be identical to the original. The background should be a clean white studio backdrop.
Add 'keep the lighting on the mug exactly as in the source photo' to your prompt.
Specify 'pure white studio backdrop with no shadows or color cast'.
Yes, write 'change the mug color to bright red while keeping the background and lighting the same'.
