Evening in Chowk AI Photo Prompt

Use the uploaded photograph(s) as the primary identity reference.

LOCK - facial STRUCTURE and body only (who they are):
- Bone structure and face shape, eye shape, eyebrow shape, nose shape, lip shape, jawline
- Skin tone, age cues, freckles/moles, natural asymmetry
- Body proportions, height cues, and build of the same person(s)
- Recognizable hair color / density (you may restyle length/part for the scene)

CHANGE for the scene (required):
- Facial expression, gaze, and head direction to fit THIS story beat
- Outfit and footwear appropriate to this specific moment
- Pose, gesture, and body language for this story beat
- Environment, props, weather, and lighting
- Hairstyle details when the scene needs it (wind, rain, festival powder, helmets, etc.)

Do not paste the uploaded outfit, pose, expression, or background into the new scene. Do not invent a new face or swap in a generic model.

Walk the uploaded person(s) slowly through a narrow old-Lucknow lane inspired by Chowk.

Aged façades, ornate balconies, old wooden doors, traditional shops, warm spill light, tangled wires, scooters, worn stone pavement, evening pedestrians. They talk quietly while walking close - natural street rhythm.

Dress in elegant understated evening wear that fits heritage Lucknow (chikankari-inspired or simple ethnic-casual) - not uploaded clothes. Pose as walking mid-stride, not standing for a portrait.

CAMERA - FULL-FRAME CINEMA STILL
Full-frame street cinema on a 35mm f/1.8 prime, slight three-quarter tracking feel. Warm shop tungsten vs cooler lane shadows. Architectural depth layers, faces crisp, background rich but not tourist-postcard sharp everywhere. Natural skin, subtle film grain.

Final check: faces and body proportions must match the upload; expression, clothing, and pose must match THIS scene, not the upload. Photoreal full-frame cinema still only - no illustration, no anime, no hand-painted cartoon look. No face swap, no de-aging, no plastic AI skin.

QUALITY: hyperrealistic skin with visible pores and natural sheen, subtle catchlights, optically clean shallow depth of field, restrained cinematic color grade, fine natural film grain - never over-smoothed AI skin or mushy phone softness.

Single photoreal frame. No collage, no grid, no duplicate person, no readable brand text or watermarks.