Use Grok Imagine on Wan 4.0 to generate a still portrait image guided by a single reference photo. Output is a still image, not video.
This route uses Grok Imagine to create still images from a text prompt and an optional reference image. Upload one reference to guide subject appearance, lighting, and mood. The output is a still image suitable for storyboarding or concept visualization.
Upload one reference image of a person. It guides facial features, hairstyle, and lighting direction.
Describe the desired portrait: expression, pose, clothing, background, and lighting. Avoid specifying resolution.
Choose from 1:1, 2:3, 3:2, 16:9, or 9:16. This ratio applies only to text-only generation. Reference edits do not forward the ratio.
A young woman with wavy brown hair and green eyes, soft smile, wearing a cream sweater. Warm golden hour light from the right. Background is a blurred autumn park. Medium close-up, eye-level composition.
The woman should have wavy brown hair and green eyes. Lighting should be warm and from the right.
Grok Imagine uses the reference as guidance, not a copy. Add more facial detail to your prompt for better match.
No. Select the ratio from the provided options. The ratio is used only for text-to-image. Reference edits ignore the ratio selection.
No. This route produces still images only. Use Wan 4.0's video routes for motion.
