聊天GPT

浪漫屋頂夜景自拍

為聊天GPT 編寫。不同型號的結果和參考照片支援可能有所不同。

使用提示詞前

列出的工具:聊天GPT。完整說明:7,193 個字元。 1 個提供的範例影像。

This prompt mentions reference images. Read its instructions for the required number of photos and identity constraints. We have not independently tested this prompt. Tool labels and supplied images are not a guarantee of compatibility or output quality.

完整提示詞

Use the uploaded user photo as the ONLY source of the primary subject’s identity. Preserve the person’s exact recognizable facial identity without alteration: facial structure and proportions, face shape, forehead, eyebrows, natural eye shape and spacing, nose, lips, cheeks, cheekbones, jawline, chin, skin tone, skin texture, hairline, hairstyle characteristics, facial hair if naturally present, and all distinctive features. Do not copy the identity or facial characteristics of anyone in the reference image. Do not beautify, reshape, slim, masculinize, feminize, age, de-age, or reinterpret the user’s face. Preserve the user’s real gender and identity exactly. The reference image controls HOW the user appears in the scene: expression, emotion, gaze, head position, pose, composition, clothing style, environment, lighting, and photographic aesthetics.
Recreate the reference as a highly realistic nighttime rooftop couple selfie, maintaining the same intimate composition and spontaneous smartphone-photography feeling. The primary user subject occupies the foreground, positioned slightly left of center and very close to the camera. Their upper body fills most of the lower and central frame. One arm extends naturally toward the lower-left foreground as if the user is holding the smartphone at arm’s length for the selfie. Keep the arm perspective, camera distance, and wide-angle selfie distortion natural and consistent with the reference. The primary subject’s shoulders and torso are angled slightly toward the camera, with the head gently tilted toward the secondary person behind them. Reproduce the reference’s relaxed, affectionate body language.
For the primary subject’s facial expression, reproduce the reference expression rather than the expression in the uploaded photo: eyes gently closed, relaxed eyelids, soft natural smile with lips lightly closed, peaceful and affectionate emotion, slightly raised cheeks, relaxed facial muscles, conveying genuine contentment and intimacy. The head should lean subtly toward the person behind them, matching the reference head tilt and rotation. Do not allow the source photo’s original expression, gaze, or head angle to influence the final expression. The user’s preserved identity must remain unmistakable beneath the transferred expression.
Behind the primary subject, include a second person positioned closely over and slightly behind the primary subject’s shoulder, creating the same affectionate couple-selfie arrangement. The second person must be a newly generated, generic individual and must NOT copy or reproduce the identity, facial structure, or distinctive features of the reference person. They should lean naturally toward the primary subject, with their head positioned above and slightly to the right of the primary subject’s head. Their expression should match the reference: a gentle relaxed smile directed toward the camera, warm and affectionate without exaggerated emotion. Their posture should feel casual and naturally close, as if both people spontaneously took the photograph together.
Recreate the clothing appearance and color palette from the reference. The primary foreground subject wears a simple fitted deep navy-blue sleeveless casual top with a subtle ribbed texture and rounded neckline, maintaining realistic fabric folds and natural fit while respecting the user’s preserved gender presentation. Add a delicate thin necklace with a small pendant around the neck. The secondary person wears a casual blue plaid/checkered button-up shirt with realistic fabric texture and natural folds, open or relaxed around the neck. Do not transfer the reference people’s identities, only the clothing concept and styling.
The setting is an outdoor rooftop or elevated terrace at night. Behind the subjects is a dense urban city skyline with numerous low- and mid-rise buildings, scattered illuminated windows, rooftop structures, trees, and distant city lights. The sky is deep dark navy-blue, with a small bright moon visible in the upper-left background. On the right side of the frame, include a rooftop railing with warm decorative string/fairy lights creating small glowing points of light. Include leafy trees and subtle rooftop details around the railing. In the lower-right background, reproduce the light-colored concrete rooftop wall/parapet containing the handwritten-style decorative phrase “Good People Brighter Days ♡” in dark lettering, positioned vertically along the wall as a subtle environmental detail.
Use realistic nighttime illumination from a combination of the smartphone/camera flash or close frontal illumination on the couple and the warm ambient rooftop lights. Faces should be clearly illuminated and naturally exposed while the background remains darker, creating the characteristic candid night-selfie contrast. Preserve realistic skin texture, natural highlights, soft shadows beneath facial features, believable flash falloff, and subtle reflections in the eyes. The moon should remain small and softly luminous rather than oversized. Warm string lights on the right should produce gentle localized glow and natural bokeh without overpowering the subjects.
Camera perspective should resemble a handheld smartphone front-camera selfie using a moderately wide-angle lens, approximately 24–28mm full-frame equivalent. Camera is held slightly below or around eye level and very close to the subjects, producing natural selfie perspective with mild wide-angle distortion. Keep the foreground arm slightly enlarged due to proximity to the lens. Maintain the square 1:1 composition and approximately the same framing: primary subject’s body and extended arm occupy the foreground, both heads remain fully visible, the secondary person rises behind the primary subject, moon appears in the upper-left, city skyline fills the middle background, and illuminated railing and concrete wall occupy the right side. Do not crop the heads, facial features, or primary subject’s extended arm unnaturally.
Create a candid, photorealistic smartphone photograph rather than a studio portrait. Preserve realistic pores, fine hair strands, natural skin variation, authentic fabric texture, believable anatomy, accurate hands and fingers, realistic depth, and subtle low-light photographic noise. Use moderate background depth of field so the couple remains sharply recognizable while distant buildings, lights, trees, and railing become slightly softer. Avoid artificial glamour retouching, plastic skin, excessive sharpening, excessive bokeh, distorted anatomy, duplicated limbs, malformed hands, unnatural facial symmetry, or an overly cinematic studio appearance.
CRITICAL PRIORITY: USER PHOTO = WHO THE PRIMARY PERSON IS. REFERENCE IMAGE = HOW THE PRIMARY PERSON appears in this scene. Transfer the reference’s affectionate expression, closed-eye smile, head tilt, gaze behavior, body pose, selfie perspective, composition, clothing aesthetic, environment, lighting, and atmosphere while preserving the user’s exact identity. The final result must look like a genuine spontaneous nighttime couple selfie taken on a rooftop, with the uploaded person unmistakably remaining themselves.

複製完整的提示並根據要求使用您自己的參考照片。

此提示提供的範例影像。此提示尚未經過我們網站的測試。

照相館
AI 產生的範例

為您自己的照片提供一點靈感。

1加入您的照片
A 人

保持臉部清晰可見

B 人

保持臉部清晰可見

JPG、PNG 或 WebP · 最大。每個 10 MB

僅當您選擇「產生」時才會發送您的照片。使用您有權使用的照片。

2讓它成為你的
更改詳情
編輯您的提示
開放雙子座

準備好…

登入您的帳戶

繼續使用您的 Google 帳戶。

您無需登入即可瀏覽和複製提示。

我們如何使用您的資訊: 隱私權政策