AI Image Prompts · 绘画提示词库

{
  "render_goal": "Contrasting fantasy portrait",
  "subject": {
    "pose": "female standing centered",
    "expression": "controlled intensity",
    "skin_tone": "fair warm glow"
  },
  "wardrobe": "{argument name="wardrobe item" default="ivory gown"} with soft {argument name="flame color" default="golden"} flames along the hem",
  "environment": {
    "location": "dark stone chamber",
    "props": "floating embers, soft shadows"
  }
}

A massive {argument name="everyday object" default="[EVERYDAY OBJECT]"} casually used as temporary shelter by tiny {argument name="travelers" default="[TRAVELERS]"}. No construction, no permanence - just rest. Overcast lighting, soft shadows, intimate mood, naturalistic storytelling.

Change the style of this collage. make it in style of {argument name="painting style" default="cezanne painting"}, and add some small nature details, but keep the whole structure of {argument name="structural elements" default="stairs and animals"} intact.

"Change her outfit style: replace the blazer and short skirt with a slightly longer Santa-inspired dress that has a pleated bottom, in classic Christmas colors (red with white accents). Add a black belt with a large buckle. Instead of the Santa hat, add a large red bow on top of her head. Keep the same pose and all other elements unchanged."

"Transform the subject of the uploaded image into a watercolor illustration rendered in a distinct, stylized aesthetic. The artwork must feature strong, expressive contour lines defining the subject's features, hair, and clothing. The color palette is limited to fluid washes of {argument name="primary color" default="teal"}, muted {argument name="secondary color" default="coral/peach"}, warm beige, and sepia tones used for shading and highlights, replacing natural colors entirely. The final image should emphasize the texture of cold-press paper, visible paint granulation, and loose watercolor bleed effects. The subject's likeness must be preserved accurately within this new medium, set against a white paper background with abstract splashes of teal and coral watercolor"

I tried to create a cute and realistic-looking girl, and initially generated it with only instructions like “{argument name="initial instruction" default="beautiful Japanese woman"},” but the AI look inevitably remained like in the first image.... So, I tested what would happen if I restructured the prompt, focusing on a subject image that the AI could easily understand, rather than just atmosphere!

Eye color, eyelashes, clothing color, length, body line.

A band of tiny {argument name="characters" default="gnomes"} throwing a colorful festival on the rim of an enormous {argument name="everyday object" default="coffee mug"}. Paper lanterns strung on wire, confetti made from magazine scraps, warm ambient city-bokeh background, shallow depth of field, cinematic wide-angle close-up, vibrant saturated palette, tactile fabric and wood textures.

An Indian folk-art depiction of a {argument name="figure" default="woman dancing"}, filled with intricate line patterns and symbolic florals, flat earthy hues, dense decorative symmetry, hand-drawn ink texture, narrative storybook folk style.

Even a "finished image" can be changed seamlessly with Nano Banana Pro!

While maintaining the impact of that design, make the background white and remove only the flashy decorations to refine it.

I am a text-based AI. Shall I provide you with code to create illustrations using Midjourney or DALL_3?