Natural portrait edits
ChatGPT Portrait Retouching Prompts
Retouch portraits with explicit constraints for identity, skin texture, hair, lighting, and expression instead of asking to “make it better.”
Before you paste
A four-step workflow
- 01Choose one correction: skin, under-eyes, flyaways, light, or color.
- 02Lock identity, expression, facial geometry, and natural texture.
- 03Request subtle strength and realistic local variation.
- 04Compare eyes, teeth, hands, jewelry, and hair against the original.
Copy and adapt
8 prompt pairs
Start short. Open the constrained version when the first result changes too much.
Reduce temporary blemishes while preserving pores and identity.Show the constrained prompt
Reduce only temporary blemishes and minor redness. Preserve the person’s identity, face shape, expression, freckles, moles, pores, fine lines, and natural skin highlights. Keep the result subtle and consistent with the existing light.Soften under-eye shadows without erasing natural structure.Show the constrained prompt
Gently reduce dark under-eye shadows while preserving natural eyelid shape, fine lines, skin texture, and facial identity. Match the surrounding skin tone and light direction; do not brighten the whites of the eyes.Tidy a few flyaway hairs while keeping the hairline natural.Show the constrained prompt
Remove only the most distracting isolated flyaway hairs. Preserve the natural hairline, curl pattern, volume, fine edge strands, and lighting. Do not create a helmet-like outline or change the hairstyle.Add soft window-light balance without changing facial features.Show the constrained prompt
Balance the portrait with soft directional window light. Lift harsh facial shadows slightly while preserving the original identity, expression, skin tone, catchlights, background, and clothing colors. Keep contrast realistic and avoid a beauty-filter look.Reduce facial shine while retaining natural highlights and texture.Show the constrained prompt
Reduce only excessive shine on the forehead, nose, and cheeks. Preserve pores, natural specular highlights, skin tone variation, makeup, and identity. Match the existing light and avoid a flat matte finish.Whiten teeth subtly to match the scene lighting.Show the constrained prompt
Make a subtle, natural correction to tooth color while preserving tooth shape, spacing, translucency, lips, and expression. Match the ambient white balance and avoid pure white, blue tint, extra sharpness, or reshaping.Create a clean professional headshot with a neutral background.Show the constrained prompt
Convert this portrait into a restrained professional headshot. Preserve identity, face shape, expression, hair, and clothing. Use even natural lighting, a neutral warm-gray background, realistic skin texture, and a crop with comfortable headroom.Correct mixed color casts while preserving accurate skin tone.Show the constrained prompt
Neutralize the mixed green and orange color casts on the portrait while preserving believable skin tone, makeup, clothing color, background mood, and identity. Keep local light variation and avoid global desaturation.See it in action
Before / after example
Real before/after examples are coming. The prompts above are structured starting points you can test on your own image.
Troubleshooting
Common failures and the next edit
The person no longer looks like themselves
Repeat that identity and facial geometry must remain unchanged; request one local correction only.
Skin looks plastic
Ask to preserve pores, fine lines, tonal variation, and natural highlights.
Eyes or teeth look unnaturally bright
Request a subtle correction matched to ambient light, with no pure white or added sharpness.
FAQ
Questions about this workflow
Can a prompt guarantee the same identity?
No. Generative edits can change facial details. State identity constraints clearly and verify the output against the source before using it.
What is the safest retouching workflow?
Make small edits one at a time, keep the original, compare at 100% zoom, and reject outputs that alter identity or important detail.