Most AI images look cool for two seconds and then you realize something’s off.
The face changes. The eyes shift. The jaw isn’t the same. It stops feeling like a real person and starts feeling like random models wearing different skins.
The three prompts below were written to avoid that problem. They lock the face in place and then let everything else change — clothes, mood, camera, and location. Instead of three random pictures, you get three versions of the same guy.
1. The KLCC Night Selfie
This prompt puts the guy in Kuala Lumpur at night with the Petronas Twin Towers glowing behind him. It’s meant to feel like a real phone selfie taken in a loud, bright city.
The messy hair, half-zipped grey jacket, black ring, and rock-on hand sign give him a modern street look without turning him into a cartoon. The low camera angle makes the towers look tall and powerful, while the flash keeps his face sharp and clear.
The important part is that his face stays exactly the same as the reference. You can change hair, outfit, and lighting, but if the structure of the face moves even a little, the whole illusion breaks. This prompt doesn’t let that happen.
Create a portrait of a young man with fair skin taking a selfie.
The face must be 100% identical to the provided reference image — no changes to facial structure, eyes, nose, lips, or jawline.
Face & Expression
His skin looks natural, clean, and healthy.
He has light brown round eyes.
His dark hair length is short and messy and volumized modern and stylish. His beard is trimmed and neat.
He wears one black ring on the index finger of his left hand.
Pose
His left hand makes a rock-on sign (🤘 shape) stylish smirk and wink while taking the selfie.
Clothing & Accessories
He wears a heather-grey zip-up jacket or sweater with a collar.
The sleeves are oversized and wide, while the body is cropped or waist-fit (semi-crop).
The front zipper is half closed.
Background
A night city view with the Petronas Twin Towers (KLCC), Kuala Lumpur, glowing brightly behind him.
Other skyscrapers are visible to the right, with subtle green lighting at the base.
The sky is deep dark night.
Lighting & Camera Angle
Strong face-focused light (phone flash or mini ring light) makes the face bright against the dark background.
The photo is taken from a low-angle, looking slightly upward to emphasize the tall towers.
Important:
The face must stay 100% identical to the reference image, transformed only into a male with messy, volumized hair and natural skin.
2. The Quiet Indoor Shot
This one is the opposite of the city scene. No skyline. No neon. Just a dim room, a soft lamp, and a close-up selfie.
The black shirt, open collar, and silver chain give him style without trying too hard. The light hitting his face is gentle, which shows off skin and bone structure better than harsh lighting ever could. His expression is calm and slightly confident, not posed.
Because the background is simple, the AI has fewer things to mess up. That makes it easier to keep the face true to the reference photo, which is the whole point.
A close-up selfie from below of a slender, well-proportioned stylish Pakistani man with a slightly upward tilt, messy Korean-style black hair, and fair skin. He has a gentle and confident expression, with slightly parted, natural pink lips and relaxed eyes. He is wearing a black button-down shirt with the top few buttons undone, exposing his collarbone and neck, and a thin silver chain necklace. The background is a dark or dimly lit room. Soft lighting from a lamp is directed at his face, emphasizing his features and hairstyle. The overall impression is stylish and effortlessly cool.
Using a real face, 100% true to the reference photo provided.
3. The Neon Street Version
Now things get loud. Fur coat, red-tinted glasses, gloves, neon signs, motion blur, fisheye lens — this one is built to look wild and high-fashion.
The stretched background and wide-angle lens make it feel like he’s moving through the city fast. The direct flash keeps his face bright while everything else fades into color and light. It’s the same guy, just turned into a street-style character.
Even with all this chaos, the face is still locked to the reference. That’s what makes it work. Without that rule, it would just be another random AI model in a cool outfit.
Create a portrait of a stylish Pakistani man with fair skin and athletic physique.
The face must be 100% identical to the provided reference image — no changes to facial structure, eye shape, nose, or jaw.
Face & Hair
He has dark, messy, volumized hair, textured and slightly wild, modern street-fashion style.
His skin looks natural, clean, and realistic — no glow filters, no blush, no cosmetics.
His eyes are light brown and round.
Outfit
Top: A thick navy fur coat (fluffy, luxurious, and bold) giving a high-fashion street look.
Accessories:
• Red-tinted rounded sunglasses with silver Frame.
• Navy gloves
Pose & Expression
He strikes a dynamic, playful pose, leaning backward or slightly spinning, with one hand reaching toward the camera (foreshortened perspective).
His hand forms a casual gesture (rock-on sign).
His expression is confident, cool, and rebellious — modern street icon energy.
Camera & Lens
Shot from a high-angle or selfie perspective using a fisheye or ultra-wide lens (0.5 style), creating slight distortion where the subject looks larger and the background curves.
Background
A busy neon-lit city street at night (Tokyo-style or another major Asian city).
Neon shop signs, street lights, and glowing boards appear blurred and stretched in the background.
Visual Effects & Lighting
• Radial motion blur in the background to give a high-speed, chaotic aesthetic
• Direct flash hitting the face, making the subject sharp and bright while the city behind fades into darkness
• Light film grain for a Y2K / retro-digital vibe
Important:
The face must stay 100% identical to the reference image, transformed only into a male with dark, messy, volumized hair and masculine fashion styling.
Conclusion
These three prompts aren’t about making three different people.
They’re about taking one real face and dropping it into three different worlds.
One is clean and urban.
One is quiet and personal.
One is bold and neon.
When the identity stays consistent, the images stop looking like AI experiments and start looking like a real person with different moods and styles. That’s what actually makes AI portraits feel believable.
Try Google Gemini Here:
This blog post and AI prompts were created by Shahbaz Ahmad.
Follow me on TikTok @Dudefrompak for more ready-to-use prompts.
📢 Join Our WhatsApp Channel
Get daily AI photo editing prompts, tools, and tips directly on your phone.
Join Now on WhatsApp 🚀
