Precise Face Swap on Red Carpet
GPT Image 2 image prompt by Elsa Ai · original post
This prompt replaces only the face and hair in a celebrity award photo while keeping every other detail identical to the target image.
the prompt
Use the uploaded TARGET IMAGE as the exact master image and the uploaded REFERENCE PHOTO only as the identity reference. Replace ONLY the main woman's face and hair with the face and hairstyle from the reference photo. Preserve her identity accurately and naturally, including facial structure, eyes, nose, lips, skin texture, and recognizable features. IMPORTANT: Do NOT copy the reference person's clothing, body, pose, background, lighting, accessories, or composition. Keep the TARGET IMAGE exactly the same in every other aspect: - Same pose and body position - Same black tuxedo - Same white pleated shirt - Same black bow tie - Same pocket square - Same pearl earrings - Same Golden Globe trophy and hand position - Same red carpet - Same photographers and background - Same “GOLDEN GLOBE AWARDS” backdrop - Same lighting, shadows, colors, camera angle, framing, depth of field, and overall cinematic look - Same image dimensions and aspect ratio Remove the small reference photo/thumbnail from the bottom-right corner completely and naturally reconstruct that area so it looks like it was never there. The final result must look like the original TARGET IMAGE, with ONLY the main woman's face and hair replaced using the uploaded reference photo. Do not make any other changes. No extra objects, no altered clothing, no changed pose, no changed background, no added text, and no unnecessary beautification or smoothing. Image: 2 Use the uploaded TARGET IMAGE as the exact master image. Use the uploaded REFERENCE PHOTO only for the person's face and identity. Replace ONLY the main man's face with the face from the reference photo. Preserve the reference person's facial identity naturally and realistically, including facial structure, eyes, nose, lips, skin texture, jawline, and beard/facial hair if present. Keep the TARGET IMAGE exactly the same in every other aspect: same hairstyle unless naturally required for the face integration, same body, same pose, same hands, same clothing, same Oscar statuette, same red carpet, same background, same photographers, same lighting, same shadows, same camera angle, same framing, same proportions, same colors, same depth of field, and same cinematic photorealistic quality. Remove the small reference-photo inset completely. Do not leave any reference image, frame, border, label, watermark, or extra text. Do NOT copy the reference person's clothes, body, pose, background, or environment. The final image must look like the same original TARGET IMAGE, with only the man's face/identity naturally replaced by the person in the uploaded reference photo. Maintain realistic skin texture and natural facial integration. Do not over-smooth, beautify, sharpen, or alter the face excessively. Final output: identical composition and dimensions to the TARGET IMAGE, with only the face identity changed.
- model
- GPT Image 2
- type
- image
- shape
- 3:4
- shared
- October 2, 2026
- prompt
- 438 words
Made this? Fix the credit or remove it.














