image-blaster is an image-to-world skillset that turns a single input image into a richer 3D production starting point. It uses Claude skills together with external generation services to create 3D environments, object meshes, Gaussian splats, and sound effects. The project is designed to accelerate early-stage 3D work by producing usable assets from visual references in just a few guided steps. It can generate dynamic object models, static environment captures, ambient loops, and object-specific physics sounds. The resulting assets can be used in game engines, 3D tools, and web-based graphics projects. image-blaster is best understood as an experimental agent workflow for quickly jumpstarting interactive worlds rather than a full 3D authoring suite.
Features
- Single-image to 3D environment workflow
- Generation of GLB and OBJ object models
- Gaussian splat output for static environments
- Ambient and object-specific sound effect generation
- Claude skill-based guided execution
- Unity, Unreal, Godot, Blender, and Three.js compatibility