Spatial visual reasoning
The cylinder, sphere, glass cube, and paper arch obey explicit front, behind, left, and right instructions without collapsing into one another.

Generate or edit detailed images with broad ratios, up to 4K quality, and as many as ten references.
The cylinder, sphere, glass cube, and paper arch obey explicit front, behind, left, and right instructions without collapsing into one another.
A three-stage diagram turns a process into a scannable left-to-right hierarchy with clear arrows and the short labels FORM, FILL, and FINISH.
The original four-object composition stays recognizable while mist, wet-stone reflections, muted surfaces, and restrained red accents adopt the atmosphere of a second reference.
Four people keep their assigned appearance, wardrobe color, prop, and left-to-right order in one detailed workshop scene.
Key capabilities
Build a simple, repeatable workflow around the inputs and controls the model reliably supports.
Capability 01
The cylinder, sphere, glass cube, and paper arch obey explicit front, behind, left, and right instructions without collapsing into one another.

Capability 02
A three-stage diagram turns a process into a scannable left-to-right hierarchy with clear arrows and the short labels FORM, FILL, and FINISH.

Capability 03
The original four-object composition stays recognizable while mist, wet-stone reflections, muted surfaces, and restrained red accents adopt the atmosphere of a second reference.



Capability 04
Four people keep their assigned appearance, wardrobe color, prop, and left-to-right order in one detailed workshop scene.


Prompt examples
These examples define subject, composition, light, and motion so you can replace the creative direction with your own.

The cylinder, sphere, glass cube, and paper arch obey explicit front, behind, left, and right instructions without collapsing into one another.
A tall blue cylinder at the back center, coral sphere directly in front, clear cube on the left, yellow paper arch on the right and slightly behind, clean tabletop, 4:3.
Try this prompt
The original four-object composition stays recognizable while mist, wet-stone reflections, muted surfaces, and restrained red accents adopt the atmosphere of a second reference.
Preserve reference 1's four-object composition exactly; transfer only the mist, wet-stone reflections, muted surfaces, ink-like edges, and red accents from reference 2.
Try this promptHow it works
Choose the input, write a clear prompt, and submit. Nanobana handles uploads, task polling, saving, and credit state.
Text or up to 10 references, broad aspect ratios, 2K, 3K, or 4K, PNG or JPEG.
Describe the subject, scene, composition, light, and details that must remain.
Review the output, then download it or continue iterating in Studio.
Campaign key art, multi-reference concepts, editorial imagery, and high-resolution deliverables.
Text or up to 10 references, broad aspect ratios, 2K, 3K, or 4K, PNG or JPEG.
Each image uses 5.5 credits.
Large outputs and many references may take longer; clear references with one visual role each work best.

Use one of the prompts above or enter your own creative direction.
Create with Seedream 5.0 Lite