Forward Future Tools Library
Spell
Spell turns a single image into an explorable 3D world, with multi-angle rendering and camera paths for visual prototyping, virtual environments, and motion design.
Try Spell →
blog.spline.design·Contact Sales
›What is Spell?
Spell is Spline's AI model for generating complete 3D scenes from an image in a few minutes. It represents scenes as volumes that can be rendered with Gaussian Splatting or other volume-rendering methods. The model supports consistent views of people, objects, environments, and 3D characters.
›What are the pros and cons of Spell?
Strengths
Generates complete 3D scenes rather than only individual assets
Maintains multi-view consistency across people, objects, environments, and characters
Supports controlled camera movement and physically motivated visual effects
Offers video, image-sequence, and volume outputs
Does not use Spline user data for training
Trade-offs
The first version still has significant room for improvement in quality and consistency
The primary scene representation is a volume, so mesh output requires a separate reconstruction technique or model
The model is still in active training, with newer checkpoints planned rather than a final fixed version
›What are Spell’s key features?
Generates 3D worlds from a single image
Renders subjects from multiple angles while maintaining scene consistency
Creates controlled camera paths
Simulates reflections, refractions, surface roughness, and depth of field
Simulates camera and object intersections instead of morphing surfaces
Exports video, image sequences, or Gaussian Splatting volumes
Allows internal volume representations to be converted into meshes with reconstruction techniques
›What are the best use cases for Spell?
Create explorable 3D environments from concept images
Produce multi-angle visualizations of objects, characters, or locations
Generate camera-path videos for film, game, and design previews
Prototype virtual reality or interactive 3D experiences
Create Gaussian Splatting assets for real-time scene exploration
›What is the pricing for Spell?
Contact Sales
›Who is Spell best for?
architectsUseful for turning reference imagery into early 3D scene and camera-view studies.
game designersWorks well for prototyping environments, characters, and interactive Gaussian Splatting scenes from images.
filmmakersA practical option for testing multi-angle shots and controlled camera paths before full production.
3D artistsWorth considering when fast image-to-world exploration matters more than production-ready mesh output.
Not for
- Teams that need consistently production-ready 3D geometry without a separate mesh-reconstruction step
- Buyers looking for a documented, mature model with settled quality and consistency
- Users who need a conventional 3D asset workflow rather than volume, image-sequence, or video outputs
›What’s new with Spell?
- Spline introduced Spell, an AI model for generating 3D worlds from images2025-01 · source
›What are the best Spell alternatives?
Hunyuan3D 2.0Choose this instead if the priority is generating high-resolution textured 3D assets.→
Instant3DChoose this instead if you want a platform focused on simplifying AI-based 3D design generation.→
Masterpiece StudioChoose this instead if you need broader AI-assisted 3D creation and modeling tools.→
Leonardo.AI→