Introducing Prompt to 3D: Generate 3D Models from Text on SphereLinks

From day one, SphereLinks has let you turn any photo into a textured GLB. Now you can skip the photo entirely. Describe what you want — “a worn leather backpack with brass buckles”, “a low-poly pine tree”, “an ornate wooden chess piece” — and our AI reconstructs it in three dimensions.

Why Text-to-3D Changes the Workflow

Image-to-3D is powerful when you have a reference photo. But for concept work, game assets, early prototypes, and procedural content, there’s often no photo to start from — there’s only an idea.

Text prompts unlock a fundamentally different use case: 3D generation at the speed of thought. No sourcing product images. No waiting for a shoot. No licensing stock photos. Just a sentence and a GLB.

• ◆

  Concept exploration: Rapidly prototype object ideas without any visual reference.

• ◆

  Game & scene assets: Generate props, set pieces, and environmental details on demand.

• ◆

  Ecommerce mock-ups: Visualise product concepts before a physical sample exists.

• ◆

  Education & research: Instantly materialise objects for interactive 3D diagrams and visualisations.

How It Works

The workflow is the same one you already know — just with a text field instead of a file picker.

1. ◆

   Describe your object. Type a clear, specific description in the prompt field. The more detail you include — material, colour, style, scale — the closer the output will match your intent.

2. ◆

   Generate. Our AI processes your prompt on dedicated GPU infrastructure and returns a fully textured 3D mesh.

3. ◆

   Refine in the editor. Open the result in the built-in 3D editor to adjust materials, transform objects, or combine the generated asset with others in your scene.

4. ◆

   Export. Download the production-ready GLB — compatible with Blender, Unity, Unreal Engine, Three.js, and every other tool in your pipeline.

Tips for Better Prompt Results

Text-to-3D is most effective when your prompt is specific. Vague inputs like “a chair” give the AI too much latitude. Specific inputs like “a mid-century modern dining chair with walnut legs and a cream fabric seat” produce far more targeted geometry and textures.

• ◆

  Name the material: “matte ceramic”, “polished chrome”, “worn brown leather”, “transparent glass”.

• ◆

  Specify the style: “low-poly”, “realistic”, “cartoon”, “sci-fi”, “medieval”.

• ◆

  Include distinctive details: Seams, logos, patterns, scratches, wear — surface detail that makes the object recognisable.

• ◆

  State the scale context: “a hand-held object” or “a room-sized structure” helps calibrate proportions.

You can also combine both modes: upload an image for rough geometry and refine the result using the 3D editor. Prompt-to-3D and image-to-3D are complementary tools in the same pipeline.

Available Now

Prompt to 3D is live for all SphereLinks accounts today. Each generation uses one credit, the same as image-to-3D. Existing credits carry over — no new purchase required.

Head to the Generate page, switch to the Prompt tab, and describe something.