@Fastmax_'s open-source Claude Code papercraft skill skips Meshy / Tripo: it unfolds a box net as UVs, has an image model paint it, and folds it back into a low-poly GLB in Blender
Fastmax (X @Fastmax_, repo author alias lejrimostfa) open-sourced Origami-3Dfolder-ai (Apache-2.0): a Claude Code skill plus Python/Blender scripts that turn papercraft-friendly props—crates, tables, windows, stalls, water towers, trees—into textured low-poly GLBs with a Fold animation. Flow: write a spec → `net.py` unfolds a UV dieline → an image model paints the template (~$0.10–0.25, with `--dry-run`) → headless Blender folds → `glb_check.py` gates. Eleven pre-painted examples ship so you can skip the paid paint. We cloned commit 4190681, installed numpy/opencv/Pillow: `net.py`+`template.py draw` produced crate template/mask; 87 author tests passed; `glb_check` on shipped `crate.glb` / `crate_fold.glb` reported ok (41 KB / 60 KB, Fold clip with 15 morphs). No Blender on the box, no paid paint call, no Claude session to invent a new asset.
Visit site Origami-3Dfolder-ai (lejrimostfa) ↗- Pricing
- Freemium
- Region
- Global
- Where it's from
- Overseas
- Added
- Nov 2, 2026
- Tested on
- Nov 2, 2026
- Version we used
- lejrimostfa/Origami-3Dfolder-ai commit 4190681; Blender not installed
- Original author
- lejrimostfa (Fastmax) @Fastmax_ ↗
- Our test
- Partly tested
- Cost / time
- Skill and scripts are Apache-2.0 free; paint via OpenAI images is about $0.10–0.25 per call (dry-run first). Blender and Python deps are local. Examples ship `painted.png`, so fold + check can be reproduced at $0.
The task
Batch cheap, consistent papercraft-style low-poly props (crates, furniture, stalls, trees) for games / Three.js without sending every asset through a general mesh generator.
Steps
- Needs Python 3.10+ (`pip install -r requirements.txt`: numpy, opencv-python-headless, Pillow), Blender 4.x (author tested 4.4), and `OPENAI_API_KEY` for the paid paint step.
- Install the skill: clone into `~/.claude/skills/origami-3dfolder-ai/`, or copy into a project's `.claude/skills/origami-3dfolder-ai/`.
- Make one folder per asset and write `spec.json` (box nets, frustum cones/cylinders, thick cards for windows/ladders; trees via `tree.py`).
- Run `python3 scripts/net.py spec.json layout.json`, then `template.py draw` and inspect `template.png` for skinny strips.
- Write `prompt.txt` (dieline rules + one line per face); `image.py edit … --dry-run` first, then drop `--dry-run` after you accept the cost.
- `template.py clean`, then `blender --background --python scripts/run_fold.py -- <asset> review/01`, finally `glb_check.py export/*.glb`. Shipped examples already have folded GLBs under `examples/*/export`.
Prompt
Use origami-3dfolder-ai to make a wooden crate: write the spec and show me the unfolded dieline first; run paint with --dry-run and wait for my OK; then export the GLB with the Fold clip.
What the author says
@Fastmax_: skip Meshy/Tripo—paint one unfolded sheet, fold in Blender, about $0.15 each (or free with shipped paints); open-source Claude Code skill. README says each object yields a low-poly GLB (25–420 KB, WEBP) plus a `*_fold.glb` with a 4 s Fold clip (15 morph targets); examples include crate, window, table, tree, Morris column, stall, handcart, water tower.
Our test
Cloned lejrimostfa/Origami-3Dfolder-ai commit 4190681; after a venv install of requirements we AST-checked all `scripts/*.py`, ran `pytest tests/test_nets.py tests/test_examples.py`, executed `net.py` + `template.py draw` on `examples/crate`, and `glb_check.py` on shipped `export/*.glb`. No Blender, no paint API call, no Claude session.
What we found when we tried it:87/87 net/example tests passed (~0.3 s). Crate unfold wrote `template.png` and `mask.png`; `glb_check` confirmed `crate.glb` ~41 KB (one mesh, WEBP) and `crate_fold.glb` ~60 KB (15 morphs, clip Fold, Blender I/O v4.4.56). So unfold + shipped GLB structure match the README; we did not run paint→Blender fold end-to-end.
How to get in
English README/SKILL; public GitHub. Default paint uses OpenAI images; Blender is a local install.
Who it suits
Indie game / Three.js builders who want a batch of consistent papercraft props and can live with boxy, conical and card shapes.
Pitfalls
- Not for organic meshes (characters, rocks)—the author says use a mesh generator there.
- Always `--dry-run` before paint; the skill must ask before a paid call.
- Needs local Blender 4.x headless; our box had no Blender, so we did not run the fold step.
- Brand-new (2026-11-01) with few stars yet; scripts and tests look complete, but it is early.
- Posted on X as @Fastmax_; CITATION/commits use lejrimostfa—credit both when citing.
More in this category
- Turn a selfie into a transparent animated sticker setImageSeedream 5.5JimengHuashu skillSkill is free; image gen uses Jimeng credits◐ Partly testedChina
- Two minutes to give agents GPT stills and GIFsImageopenai-imageMCPGrok BotMCP free; images use OpenAI credits◐ Partly testedOverseas
- One sentence to batch WebP and trim edgesImageimage-processingClaude CodePillowSkill free; needs Claude Code + Pillow◐ Partly testedOverseas
- One photo plus a minimal paper-print twinImageCodexSkillImage genSkill free; needs Codex image generation◐ Partly testedChina
- Pixel VFX as code—plus a Claude refine loopImagepixelfxClaude Codepixel VFXLib+skill free; new FX need Claude◐ Partly testedOverseas