Image Blaster — Image-to-3D World Skillset for Claude
Creates 3D environments, SFX, and meshes from a single image using Claude skills, World Labs, and FAL. Takes you from an image to a fully meshed 3D environment in under 5 minutes — generates 3D models (.glb, .obj), Gaussian splats (.spz), and ambient sound effects (.mp3).
Prompt template
Put an image into input/ directory and ask Claude to blast it and confirm each step with me.
How to use
- 1Clone the repo — git clone https://github.com/neilsonnn/image-blaster && cd image-blaster
- 2Install Claude CLI — curl -fsSL https://claude.ai/install.sh | bash, then run claude in the project directory
- 3Provide API keys — give Claude your World Labs API key and FAL API key (for Hunyuan 3D and ElevenLabs SFX)
- 4Prepare input image — place your source image into the input/ directory
- 5Trigger the blast — ask Claude to 'blast it and confirm each step with me' to start the image-to-3D pipeline
- 6Environment generation — World Labs Marble model (marble-1.1) creates an explorable 3D environment from the input image
- 7Image cleanup — nano-banana (or gpt-image-2 as alternate) handles source cleanup, clean plates, and object reference images
- 83D model creation — Hunyuan 3D model via FAL generates .glb and .obj models of all dynamic objects, with configurable face count (default 50000), PBR materials, and generation type (Normal/LowPoly/Geometry)
- 9Gaussian splat generation — creates a .spz Gaussian splat file of the static environment for high-fidelity rendering
- 10Sound effects generation — ElevenLabs SFX model creates ambient looping sound and object-specific physics sound effects as .mp3 files
- 11Export and embed — use the generated assets in Unity, Unreal, or Godot game engines; Blender, 3DS Max, or Maya DCC software; or Three.js web apps and Electron apps
Pro tips
- Go from image to fully meshed 3D environment in under 5 minutes — great for jumpstarting 3D work
- Hunyuan 3D parameters: --face-count 40000-1500000 (default 50000), --enable-pbr true/false (default true), --generate-type Normal/LowPoly/Geometry, --polygon-type triangle/quadrilateral
- Perfect for video game level concepts, architectural renderings, film location scouting, robot environments, and recreating childhood bedrooms
- Remove /app from .claudeignore to give Claude the ability to modify the built-in React viewer
- The skillset separates dynamic objects (3D meshes) from static environment (Gaussian splat) for optimal rendering performance
- MIT licensed — embed image-blaster assets under any game engine, DCC software, or web app
- Works with Claude CLI natively; the .claude/ directory contains the skill definitions that Claude auto-discovers
Works best with
Ratings & Reviews
Loading reviews…
Related Skills
Emil Kowalski Skills — AI Skills for Designers & Engineers
A curated collection of 12 AI skills for designers and engineers, built by Emil Kowalski (ex-Vercel, ex-Linear). Helps agents make great animation and design decisions — from building animations from scratch to auditing your codebase, picking the right UI library, and writing modern Swift. 45.1k GitHub stars.
Ecommerce Image Skills — 12 Agent Skills for Product Photography
Twelve installable Agent Skills that turn a casual product photo into platform-ready ecommerce images — Amazon main images, lifestyle scenes, infographics, A+ modules, size charts, on-model shots, sale banners, Xiaohongshu covers and Taobao detail pages. Powered by GPT Image 2.5 (Flare / Sunburst).
Oil UI — Push AI's UI Design Capability to the Limit
A structured AI UI design skill framework with an 8-step design methodology, style comparison pages, independent review, and automatic updates. Generate multiple distinct design directions, compare them side by side, and polish interfaces to production quality.