Dark Seinen: Lone Wolf Protagonist at Dusk

Anime illustration, dark seinen aesthetic, lone wolf protagonist, [your selfie] as a brooding anti-hero standing at the edge of a rain-soaked rooftop at dusk, city skyline glowing amber and crimson far below, long dark overcoat billowing in the wind, one hand in pocket, other hand loosely gripping a katana at the side, face partially obscured by windswept dark hair, sharp angular jawline, tired determined eyes with faint dark circles, moody atmospheric lighting, neon signs reflecting on wet concrete, low dense clouds, color palette of deep charcoal, burnt orange, blood red, and cold blue highlights, Berserk and Vinland Saga art direction, extreme detail in fabric folds and rain droplets on surfaces, cinematic wide shot with subject on the left third, negative space on the right, vertical portrait orientation, masterpiece quality anime illustration

Why it goes viral

The dark anti-hero archetype + Berserk/Vinland Saga aesthetic taps the most passionate corner of anime fandom, seinen fans save and share detailed atmospheric art obsessively.

How to use it

Upload your selfie to Flux 1.1 and set it as the img2img reference at 65% denoising strength. Paste the full prompt into the positive prompt field and add the suggested negative prompt. Set resolution to 768x1344 and run 30-40 inference steps for maximum detail. Upscale the best result with RealESRGAN Anime 4x to bring out rain droplets and fabric texture. Post with a moody one-line caption in Japanese or a lyric from an anime OST to match the aesthetic authenticity.

Pro tips

In Flux 1.1, use the img2img workflow at 60-70% denoising strength with your selfie as the base to maintain facial likeness while heavily stylizing. The charcoal + burnt orange + blood red palette is critical, specify hex values (#1a1a1a,

Steps and settings

1. 01Upload your selfie to Flux 1.1 and set it as the img2img reference at 65% denoising strength.
2. 02Paste the full prompt into the positive prompt field and add the suggested negative prompt.
3. 03Set resolution to 768x1344 and run 30-40 inference steps for maximum detail.
4. 04Upscale the best result with RealESRGAN Anime 4x to bring out rain droplets and fabric texture.
5. 05Post with a moody one-line caption in Japanese or a lyric from an anime OST to match the aesthetic authenticity.