ChatGPT

Magical Lantern Night — Cinematic Portrait Collage

Written for ChatGPT. Results and reference-photo support may differ across models.

The complete prompt

Use the uploaded user photo as the ONLY identity reference for the person in this image. Preserve the user's exact recognizable identity with maximum fidelity: facial structure, face shape, forehead, eyebrows, natural eye shape, nose, lips, jawline, cheekbones, chin, skin tone, natural skin texture, hairline, hair characteristics, and all distinctive facial features. Do not replace the user's identity with the woman shown in the reference. Do not beautify, slim, reshape, feminize, masculinize, age, de-age, or otherwise reinterpret the user's face. The user's identity must remain unmistakably recognizable in every visible panel. The reference image controls everything except identity, including expression, gaze, head position, body pose, clothing, composition, lighting, environment, and photographic style.
Recreate the reference as a single vertically stacked three-panel cinematic photographic collage with three horizontal scenes separated by very thin clean white divider lines. Maintain the same overall tall portrait aspect ratio and panel proportions. The setting is a magical nighttime rooftop overlooking a city lake/waterfront, filled with hundreds of glowing orange paper sky lanterns floating throughout a deep navy-blue night sky. A bright full moon is partially surrounded by soft dark clouds in the upper-right sky. The distant city skyline glows with warm golden and scattered colorful lights, with reflections shimmering across the water. Rooftop edges, plants, flowering bushes, decorative lanterns, and warm architectural lights create a romantic festive atmosphere. Use realistic nighttime exposure, warm orange lantern illumination against cool blue surroundings, cinematic depth, natural skin rendering, realistic hands, physically accurate lighting, subtle bokeh, and premium photorealistic photography.
TOP PANEL — Wide rear-view cinematic shot. Show the user's identity-bearing subject from behind, standing near a rooftop terrace and facing the lake and city. The subject occupies the lower-center portion of the frame, with long natural hair falling down the back if compatible with the user's identity. They wear a flowing elegant black traditional South Asian outfit with delicate small floral embroidery and a sheer/flowing black dupatta. Both arms are raised naturally above shoulder level, holding and releasing a large glowing orange paper sky lantern directly overhead. Reproduce the exact hand and arm positioning: one hand supports the lower side of the lantern while the other reaches toward its opposite edge, creating a natural upward-reaching pose. The lantern is large, warm orange, softly illuminated from its small flame inside, with visible thin paper construction and gentle translucency. Fill the expansive sky with dozens upon dozens of smaller glowing lanterns at different distances and scales, creating strong atmospheric depth. Place the moon and clouds in the upper-right quadrant. Include flowering plants and dark foliage along the lower-left edge, a warm decorative rooftop lantern on the lower-right, and the illuminated waterfront city across the bottom background. Camera is positioned behind the subject at approximately standing eye/chest height, using a wide cinematic perspective and deep environmental composition. Do not show the user's source-photo expression in this panel because the face is turned away.
MIDDLE PANEL — Intimate close-up portrait. Show the same user's identity clearly and accurately, now matching the reference subject's warm, gentle, happy expression: a subtle genuine smile with relaxed cheeks and naturally bright eyes. The head is tilted slightly upward and toward the upper-right area of the frame, with the eyes looking upward toward the glowing lantern rather than directly at the camera. Preserve the reference's soft, dreamy emotional mood. Frame from approximately upper torso/chest upward, with the face occupying the left/center portion while the large glowing paper lantern fills the upper-right foreground. Both arms are raised toward the lantern; one hand reaches underneath/along its lower edge and the other hand holds the opposite side. Reproduce the natural hand anatomy and graceful arm positioning. If the user's identity has different hair length or texture, preserve their real hair identity while arranging it naturally to match the reference framing. Dress the subject in the same elegant black embroidered traditional outfit, with delicate floral motifs, sheer black fabric, a black choker-style necklace, small earrings, and stacked metallic bangles on the wrist, adapting accessories naturally without altering identity. The lantern casts strong warm amber light across the face, fingers, hair, and clothing, while the background remains dark and softly blurred. Add numerous circular orange lantern bokeh lights behind the subject. Use shallow depth of field, realistic skin texture, cinematic warm highlights, soft falloff, natural catchlights, and a premium editorial night-photography aesthetic. The user's original facial expression must be completely replaced by the reference expression.
BOTTOM PANEL — Medium/full seated rooftop portrait. Show the same user's identity seated comfortably on a rooftop ledge or low terrace structure, surrounded by the same nighttime waterfront environment and floating lanterns. Match the reference pose precisely: body angled slightly toward the camera, one elbow resting naturally while the corresponding hand supports the side of the face/chin, with the other arm relaxed across the lap. The head is gently tilted and turned toward the upper-left side of the image. The eyes look softly upward and slightly to the left, away from the camera, matching the reference gaze. Expression is peaceful, dreamy, affectionate, and subtly smiling rather than neutral. Preserve the user's exact identity while reconstructing this expression and gaze. The subject wears the same flowing black South Asian traditional outfit with delicate light-colored floral embroidery, sheer sleeves/dupatta, black choker necklace, small earrings, and metallic bangles. Keep the clothing elegant and naturally draped around the seated body. Position the subject approximately center-right/lower-center, with a warm decorative metal lantern glowing brightly in the lower-right foreground. Include flowering plants and foliage on the lower-left, a low rooftop parapet, distant illuminated buildings, water reflections, and a sky filled with floating orange lanterns behind the subject. Use cinematic shallow-to-moderate depth of field, warm rim lighting from nearby lanterns, soft orange highlights on the hair and clothing, cool navy ambient sky light, realistic shadows, and beautiful natural bokeh.
Maintain visual continuity across all three panels: the exact same user identity, consistent facial proportions, consistent skin tone, consistent hair identity, consistent outfit, consistent accessories, and consistent nighttime location. The three panels should feel like different photographs from the same magical lantern festival moment. Photorealistic, ultra-detailed, cinematic, realistic anatomy, realistic hands and fingers, natural fabric texture, physically accurate lantern glow, atmospheric perspective, subtle filmic contrast, high dynamic range, refined color grading, detailed city lights and water reflections, soft background bokeh, premium professional photography, 8K-quality detail. Do not introduce extra people, text, logos, watermarks, borders other than the thin white panel dividers, distorted anatomy, duplicate limbs, malformed hands, artificial facial features, excessive skin smoothing, or identity drift. Most important: USER PHOTO = WHO THE PERSON IS; REFERENCE IMAGE = HOW THE PERSON LOOKS, POSES, EXPRESSES, GAZES, DRESSES, AND INTERACTS WITH THE SCENE.

Copy the complete prompt and use your own reference photos where requested.

Example image provided with this prompt. This prompt has not been tested by our site.

Photo studio
AI-generated example

A little inspiration for your own photo.

1Add your photos
Person A

Keep the face clearly visible

Person B

Keep the face clearly visible

JPG, PNG or WebP · Max. 10 MB each

Your photos are sent only when you choose Generate. Use photos you have permission to use.

2Make it yours
Change the details
Edit your prompt
Open Gemini

Getting ready…

Sign in to your account

Continue with your Google account.

You can browse and copy prompts without signing in.

How we use your information: Privacy policy