Alici AI

Explore. Create. Go viral.

Browse trending looks, switch between featured topics, and open standout prompts or ready-to-use scenes without leaving the page.

Describe what you want to create...
[Subject] A side-by-side AI model comparison image showing the same young woman rendered in two nearly identical portraits. She appears early 20s, feminine presentation, slim build, light-medium skin tone, long dark hair in a high ponytail with loose face-framing strands, large round silver wire-frame glasses, medium hoop earrings, and a black sleeveless scoop-neck top. Her facial expression is the key feature: worried, sad, and slightly teary-looking, with knitted brows, downturned mouth, and vulnerable eyes looking toward the camera or slightly off-center. One hand rests on a desk surface near the lower edge of the frame. [Environment] Indoor room with cool blue ambient lighting, modern home-office or bedroom workspace vibe, desk edge visible at the bottom, dark chair back behind, framed wall art in the background, and a glowing blue neon sign partially visible on the right side. [Composition/Camera] Two vertical portrait panels placed side by side with narrow spacing, each framed from upper chest to above head, subject centered in each panel, consistent angle and crop for direct comparison. Bottom labels identify the models, with one panel reading SEEDREAM 4 and the other NANO-BANANA. Small prompt text near the center-bottom reads in Spanish: La chica esta triste. [Lighting] Cool indoor ambient blue lighting with soft frontal face illumination, low-to-medium contrast, smooth skin exposure, moody but readable emotional portrait lighting, no harsh flash. [Style/Rendering] Realistic AI comparison-cover graphic, clean high-detail portrait rendering, emphasis on micro-expression fidelity, polished skin, slight background blur, comparison-thumbnail aesthetic. [Detail constraints] Keep the layout as a two-panel side-by-side comparison of the same woman, preserve the sad worried expression, glasses, black top, blue-lit interior, desk hand placement, and model labels; do not add extra people, costume elements, bright outdoor light, or mismatched outfits between panels. Negative prompt: smiling, laughing, aggressive anger, extra people, different hairstyles between panels, missing glasses, cluttered office, colorful wardrobe, no desk, no neon sign, distorted hands, duplicate limbs, uneven crops, cartoon style. Suggested parameters: aspect ratio 4:5 overall cover, lens 70-85mm equivalent portrait feel, shallow depth of field, 30-40 steps, CFG 6.5-7.5, sampler DPM++ 2M Karras, seed 734118. Delta prompt strategy: 1) If the expression is not sad enough, append 'worried sad expression, knitted brows, downturned lips, vulnerable eyes'. 2) If the panels diverge too much, append 'same woman, same crop, same styling in both panels for direct comparison'. 3) If glasses disappear, append 'large round silver eyeglasses clearly visible in both panels'. 4) If the room loses its mood, append 'cool blue ambient indoor lighting with a partial neon sign on the right'. 5) If the top changes, append 'black sleeveless scoop-neck top'. 6) If the hand vanishes, append 'one hand resting on the desk at the bottom edge'. 7) If the crop is wrong, append 'upper-chest portrait crop in both side-by-side panels'. 8) If the background gets too busy, append 'minimal blue-lit workspace background with chair and framed wall art'. 9) If labels disappear, append 'bottom panel labels naming each model for comparison'. 10) If the look turns too glamorous, append 'natural emotional portrait focused on expression fidelity'.
Generate โ†’

Trending Now

๐Ÿ”ฅ Seedance 2 is here

Let's Dance.

Born to go viral.

History called. We picked up.

Jujutsu Kaisen meets your world.

Not real. Still iconic.

Featured Creators

Browse All