
Original document
GPT Image 2 · zh
This is a dark, cinematic, Chinese ink-wash fantasy illustration. At its center is a large black calligraphic Chinese character, "安" (An), written with bold, powerful strokes on a white, luminous, full-moon-like background. In the lower right foreground, a scholar-warrior from East Asia is depicted, turned to the left, with disheveled hair that flutters in the wind. His robe is tattered, resembling smoke and ink stains. The figure stands solemnly, almost ghostly, amidst a storm of mist and swirling black ink clouds. The color palette is dominated by charcoal black, ashen gray, and cool white, accented only with touches of deep red. The illustration includes four visible pagodas: two small pagodas in the distance on the right, one reflected in a large, shattered mirror in the lower left, and a taller pagoda emerging from the mist in the upper right. The composition is surrounded by five striking shards of broken glass or mirror: a large triangular shard in the lower left, a towering shard near the center at the top, a large vertical shard in the upper right with a smaller, suspended shard below it, and a large, curved shard in the lower right, along with a few smaller fragments. One of the shards on the left contains a partial reflection of a human face. A red seal bearing Chinese characters is added to the left below the large calligraphic characters. The use of splattered ink textures, torn paper edges, dripping red marks, and layered smoke creates an eerie, martial arts-themed atmosphere. High-contrast lighting, a mixed-media painting style, textured brushstrokes, a poster-like composition, extremely high detail, a strong sense of atmosphere, and a mysterious yet elegant feel.How to use
- Copy the complete original prompt, including any parameters.
- Replace the subject, text and composition with your own requirements.
- Use a compatible external tool; check its current access conditions and review the result.
Before you start
- Check whether a reference image is required. A compatible model label does not mean this example was generated with that model.
- Keep names, figures and visible text accurate; results can vary between model versions.





