
A highly specific JSON prompt for generating an ultra-realistic, cinematic portrait of Ana de Armas in a squatting pose within an urban parking garage. Crucially, this prompt uses 'identity lock' and 'pose lock' rules, requiring the AI to strictly adhere to external reference images for the face, hair, pose, scene, and lighting, while applying a modern street fashion outfit.
{
"scene_description": {
"environment": "Indoor parking garage, concrete floor, industrial ceiling pipes, columns, urban infrastructure architecture",
"background": "Wide and empty parking area, perspective depth, industrial atmosphere",
"floor": "Asphalt/concrete floor, natural texture, slight reflections",
"mood": "urban, cinematic, minimal, modern"
},
"subject": {
"gender": "female",
"identity_source": "reference_1",
"identity_lock": {
"face": true,
"hair_color": true,
"hair_texture": true,
"hair_style": true,
"hair_density": true,
"hairline": true
},
"pose": {
"body_position": "Woman in a squatting position (squat pose), heels on the ground, knees bent",
"posture": "Balanced, aesthetic squat stance",
"head_pose": "Head slightly tilted to the side",
"gaze": "Direct eye contact with the camera",
"expression": "Calm, cool, slightly mysterious facial expression"
},
"hands": {
"hand_position": "One hand resting on the chin, the other hand in a relaxed position"
}
},
"outfit": {
"top": "Oversized light-colored sweatshirt",
"bottom": "Short skirt 





















































