Spell

Spell is an experimental AI product from Spline that turns a text prompt or uploaded image into an interactive 3D scene, which Spline calls a “World.” It is designed for generating a scene concept from a single visual input and sharing the result through an embeddable public link.

What it does

Spell supports two starting workflows. A text prompt generates four related images, while an uploaded image returns the original alongside three variations. After selecting an image, users can use the “Generate 3D” control to start scene generation. Completed scenes are added to a queue, and Spell provides a notification when a scene is ready.

Spline documents the output as a volumetric representation rendered with Gaussian Splatting. The company describes its underlying diffusion model as designed to maintain consistency across multiple viewpoints for subjects such as people, objects, environments, and 3D characters. It also says the system can render different camera views and generate controlled camera paths.

Notable capabilities

• Generate image variations from either a text prompt or uploaded image.
• Convert a selected image into a 3D scene through the “Generate 3D” workflow.
• View completed scenes interactively with controls for background, auto-orbit, orbit, pan, zoom, touch behavior, and interaction limits.
• Share scenes through a public link that can be embedded on the web.

Who it helps

Spell is suited to creators, designers, and teams exploring 3D concepts from reference images without starting from a conventional 3D modeling workflow. It may also fit presentations or web-based prototypes where an interactive scene and controlled camera movement are useful. Because Spline describes the product as highly experimental with limited access, it is better treated as an exploratory generation tool than as a guaranteed production pipeline.

How it fits a workflow

A typical workflow begins with a prompt or image, continues through image selection and 3D generation, and ends with interactive review and link-based sharing. Users can configure how viewers move through the scene before embedding or distributing it. The service’s pricing page lists a free plan with access to the inspiration library but no scene-generation access, alongside an Early Supporter plan priced at $99 per month with scene generation, the library, and watermark-free export.

Strengths and limits

Spell’s main strength is its short path from a single image or text idea to a viewable, shareable 3D scene. Its multi-view focus may be useful when a concept needs to be inspected from more than one angle. Spline cautions that results can be unexpected, that well-defined shapes work better, and that abstract or overly complex inputs may produce poor results. Access is limited, and the company says early-stage pricing may change.