AI Image Generation

SphereLinks includes an AI image generator that creates reference images from text prompts. Use it when you don't have a photo of the object you want to turn into 3D.

How it works

The generator uses Claude's code execution capability to render a clean, centred illustration of your subject on a white background. The output is a 512 × 512 PNG optimised for 3D generation.

Using the generator

1. On the Generate page, click Generate from prompt (the pencil icon) instead of uploading a file.

2. Type a description of your object, e.g. "a terracotta flower pot with a small succulent".

3. Click Generate Image. The AI renders the image in a few seconds and loads it into the upload zone.

4. Review the preview. If it looks right, proceed with Generate 3D Model. If not, adjust the prompt and regenerate.

Writing effective prompts

• Name the specific object, not a scene: "a wooden rocking chair" not "a cosy living room".

• Add material or style cues: "polished chrome", "weathered oak", "cartoon style".

• Avoid multiple objects in one prompt — generate them separately and combine in the editor.

Image generation is free and does not consume credits. Only the subsequent 3D generation step costs 1 credit.