wan2.7-image Portrait Style Comparison: Realistic vs. Artistic

This work is aimed at developers and design teams who need a complete set of portrait visuals. It demonstrates how to use the same reference portrait with different prompts to obtain two style categories of output—realistic and artistic—on wan2.7-image-standard. In the schema, prompt is a required field, and image_urls can accept up to 9 reference images; once reference images are passed in, the output aspect ratio will follow the last reference image. The resolution defaults to 2K, and aspect_ratio only takes effect for pure text-to-image generation. Therefore, this entry first uses one reference portrait to lock in the subject's bone structure and styling, then uses one complete prompt to describe the lighting, materials, and mood of the image. This is a typical combined usage of reference image + text. It is suitable for early-stage style exploration for virtual models, character design, and brand portrait materials, and can also serve as a comparison sample for prompt writing. The work was genuinely generated by the author's own pipeline.