Magic Poser Web
Magic Poser is a browser pose base for character posing—set 3D figure drawing references, adjust joints and camera angles, and use the embed from webapp.magicposer.com.
Turn a reference photo into editable, animation-ready Three.js TypeScript—browse the img2threejs converter and procedural gallery.

Procedural Three.js output: Get typed factories and structured specs you can version, review, and wire into a scene—hard-surface objects, characters, and dioramas from a single reference.
Native Mesh export option: When you need a textured asset for a downstream 3D pipeline, Native Mesh mode produces GLB or glTF instead of editable code.
Reference photo side-by-side: Open demos to see the photo each model was reconstructed from, so you can evaluate silhouette, materials, and animation readiness.
Orbit, zoom, and inspect: Review pivots, idle loops, and part hierarchies that make the model animation-ready rather than a static lump.
Agent skill + hosted workflow: Install the open-source skill for Claude Code, Codex, or OpenCode, or use the hosted converter on img2threejs.org for credit-based generation.
Browse the gallery or converter
Open img2threejs.org to scan showcase subjects—cameras, tools, controllers, appliances—and judge fidelity before generating.
Choose Procedural or Native Mesh
Pick Procedural for editable Three.js TypeScript drafts, or Native Mesh when you need a textured GLB/glTF asset.
Prepare clear reference evidence
Procedural mode works best with a clean front/side/top board; Native Mesh needs one well-lit image with the full subject visible.
Generate, preview, and export
Run the hosted flow or the agent skill, review photo-versus-render comparisons, then export TypeScript/JSON or mesh files into your project.
Creative coding and Three.js apps
Land maintainable scene code with named parts, pivots, and materials you can refactor like application software.
Product and prop visualization
Show stakeholders animation-ready 3D from a still before committing to a full mesh or CAD pipeline.
Game and simulation prototypes
Move from a photo to structured hierarchy with addressable parts for controls, collisions, and animation.
Agent workflow evaluation
Decide whether a code-first image-to-3D skill fits your stack versus downloading textured meshes from other tools.
Freemium hosted converter: img2threejs.org offers credit-based Procedural drafts and Native Mesh exports; confirm current pricing and limits on the official site.
Open-source skill: The MIT-licensed upstream skill lives at github.com/img2threejs/img2threejs for Claude Code, Codex, and OpenCode installs.
Best input types: Hard-surface subjects with clear silhouettes work strongest; Procedural prefers three-view boards, Native Mesh prefers one clear normal photo.
Page type: This route is a curated listing at /img2threejs. Conversion, credits, and exports happen on img2threejs.org.