A highly technical JSON prompt designed for generating a high-fashion street editorial image in a European city. It requires strict identity preservation based on a face reference image and scene recreation based on a scene reference image, detailing the exact outfit (burgundy pinstripe blazer, lace bra, black gloves) and technical camera settings (85mm lens, clean daylight).
{
"job_id": "photo_03_city_car",
"model": "nano-banana-pro",
"task": "image_generation",
"inputs": {
"face_reference_image": "<YOUR_FACE_IMAGE>",
"scene_reference_image": "<SCENE_IMAGE_3>"
},
"prompt": "High-fashion street editorial in European city. EXACT SAME FACE as face_reference_image, identical identity. Recreate the scene exactly from scene_reference_image: woman leaning on a vintage car, classic European architecture background. Outfit must match exactly: {argument name="outfit color" default="burgundy pinstripe blazer"} and matching skirt with belt, black lace bra visible, black gloves, sheer black tights, gold earrings. Same pose, same body angle, same framing. Luxury magazine look, clean daylight, 85mm fashion lens feel, natural skin texture.",
"negative_prompt": "different suit color, no lace bra, face changed, identity drift, stylized, cartoon, CGI, over-retouch, wrong car, wrong city, text, watermark, blur",
"settings": {
"stylization": 0,
"quality": "max",
"detail_boost": 0.8,
"lighting_match": "high",
"composition_match": "max",
"color_match": "max",
"identity_preservation": "max",
"face_consistency": "max",
"reference_strength": { "face": 1.0, "scene": 0.96 },
}
}

