"Make people love each other and be polarized like never before at the same time. Add this exact prompt to the image"

{ "input_reference": "@Image1", "visual_style": "Casual home shot with warm indoor lighting, slight camera shake", "subject": "Framed with a male African lion (possessing a thick mane)", "setting": "Lion's nest sofa, lion's head res

{ "task": "image-to-image restoration", "input_image": "output_from_prompt_1", "reference_image": { "type": "face_reference", "description": "Same face as damaged photo, identity must remain unchanged" }, "restoration_settings": { "restore_

@Image, stunning Asian beauty, tall and slender figure, 10-second sweet and cool style rapid outfit change, smooth action transitions; 0-1s: Academy uniform, hands behind back tiptoe jump, tie swinging, tilting head with eyes smiling like c

{ "prompt": { "subject_description": { "main_subject": "Portrait of a stunning young woman with long, voluminous dark brunette hair", "facial_features": "Hazel-brown eyes with sharp winged eyeliner, glowing dewy skin textu

{ "image": { "type": "photograph", "setting": "brightly lit gym with mirrors and weight machines", "subjects": [ { "name": "Victoria Justice", "position": "left", "attire": { "top": "dark grey sports bra", "bottom": "dark grey workout short

{ "meta": { "quality": "ultra photorealistic", "resolution": "8k", "camera": "iPhone 15 Pro", "lens": "26mm", "aspect_ratio": "4:3", "style": "warm lamp, grainy iphone look" }, "character_lock": { "age": "mid 20s", "ethnicity": "blonde euro

Create a complete visual worldbuilding model for a stylistic civilization based on reference image @image_1, including multiple images of architecture, characters, costumes, vehicles, and maps, adopting a unified design language, with cinem

{ "generation_request": { "meta_data": { "task_type": "pastel_meadow_fashion_editorial", "priority": "highest", "language": "en", "style_version": "v1.0_summer_meadow_pink_dress_flash_haze" }, "output": { "aspect_ratio": "4:5", "resolution"

15 seconds, 9:16 vertical screen, deliberately mimicking low-quality mobile phone video recording, handheld POV, high noise + motion blur + digital compression distortion + timestamp watermark (top left corner, 2007 style), 0-5 seconds: Han