0 / 20000
Create illustrations, product concepts, and visual references with the AI image generator on 3D Design AI. Start from text or an existing image, choose the available aspect ratio and resolution, and download artwork for your layout or 3D project.
Choose your starting point based on how much of the composition you have already decided.
Describe a subject, its surroundings, and the visual treatment you want. This is useful for exploring an idea before you have artwork, whether it is an editorial illustration, an object concept, or a scene for a presentation.
Upload a reference and describe the change you want, such as a different setting, material, or style. State which features should stay recognizable. An AI-generated variation may also change details you intended to keep, so compare it with the original.
A useful brief explains what matters in the picture. Add detail where it affects the result.
Name the main subject and explain where important objects sit relative to it. For a crowded composition, establish the focal point before adding secondary elements.
Describe the viewpoint, crop, and background. Leave room for a headline or product label if you plan to add typography later. Choose an available aspect ratio that matches the intended layout.
Specify the lighting direction and the surfaces that matter: matte, polished, translucent, or textured. Consistent visual instructions are easier to assess than several conflicting style requests.
Compare each result with your intended use before choosing the next change.
Choose an aspect ratio for the layout: a square composition for an isolated object study, portrait framing for a character, or a wide view for a scene. Set the output resolution offered by your model once the composition works.
For several references, Nano Banana Pro accepts up to 8 images here, while Nano Banana 2 accepts up to 14. Explain which image supplies the subject, style, or setting, and keep the references consistent with the same brief.
Compare the result with the features that define your subject: facial features, product markings, object count, and repeated patterns. Keep the version that best captures your intended design.
In image-to-image mode, name the specific change and what should remain. Compare the revised result against both your brief and the reference. Keep a promising version before trying a substantially different direction.
Use the finished image on its own or prepare it for another creative tool.
Generate one complete object with a readable silhouette and minimal overlap. Inspect the result for inconsistent parts before using it as a 3D reference. This page produces a flat image; a separate generation step creates the 3D model.
Create a 3D model from your imageFor an existing image that only needs its background removed, use the dedicated cutout workflow. To create a new subject on transparency, start with the transparent image generator.
Answers about creating, editing, and choosing your image output.
Text-to-image starts from your written description. Image-to-image also uses uploaded visual material to guide a new result. Choose text for an open-ended idea and a reference when there is a composition or subject you want the generator to work from.
Start with the input you have: text for a new scene, or image-to-image for an existing subject. For multiple references, compare the upload limit; Nano Banana Pro and Nano Banana 2 both support multi-image input here. Then choose a model with the aspect ratio and resolution your layout needs.
You can request lettering as part of the image, but check every word and character in the result. Short, clearly specified text is easier to review. For a final logo, exact label, or text that must remain editable, add the typography in your design application after generation.
No. Available aspect ratios, resolutions, and reference limits depend on the selected model. Use the options shown in the generator, and check the downloaded image dimensions before placing it in a layout that requires a particular size.
A reference guides the result but does not guarantee an exact copy. Review proportions, markings, facial features, and small details after editing. If you only need to remove the background while retaining the existing subject, choose background removal instead of generative editing.
No. A picture with realistic depth, shading, or a rendered appearance is still a flat image. To obtain a model you can rotate and inspect in 3D, use the AI 3D model generator with a suitable reference image.
Yes. Limited free image generation is available on 3D Design AI. Sign in to see which options you can use and create an image from your own brief.
Describe the scene or upload a reference, then refine the result around the details that matter to you.
Create an Image