3D models from photos and text online — AI

Bottom Line Up Front (BLUF): The 3D section gathers every way to get a 3D model on NeuralSpace: from a single photo, from a text description, from multiple views, and for free right in the browser. Every tool has its own address, its own price and its own input requirements.

Which tool to choose

Photo → 3D (TRELLIS-2) is the fastest way from a snapshot of an object to a textured model. Text → 3D (Tripo) is for when the object does not exist yet and you only have a description. Multiview → 3D is for accurate sides and back: up to four views of one object are uploaded.

The free tools run on your device’s graphics card via WebGPU: the photo never leaves the browser and no tokens are charged. Paid cloud runs charge tokens when the task is created and return them on failure.

What you get

The platform returns a GLB file with geometry, textures and material maps, which opens in Blender, Unity, Unreal Engine and most 3D editors. The “Photo → 3D splat” tool returns a 3D Gaussian splat in .splat format for splat viewers.

An example before you generate

The “Show example” button opens a finished textured model: spin it and inspect it from every side without spending tokens.

Frequently asked questions

Is there a free way to make a 3D model?

Yes: “Photo → 3D (free)” on TripoSR and the experimental splat tool run on your device via WebGPU and charge no tokens.

How are paid runs different from free ones?

Paid models run in the cloud: the result is more accurate and does not depend on your hardware, but each task costs tokens. Free tools run in the browser and need a recent Chrome or Edge with WebGPU.

What if generation fails?

Tokens for a failed task are returned automatically. Try another input: one object on a plain background in good light for photos, or a more detailed description for text.