OpenSKP

The open-source SketchUp (.skp) file parser — Python edition.

Parse .skp files without SketchUp. No SDK. No license. Just code.

🏠 openskp.com · 🌐 Try the Live Web Viewer · 📖 Docs · Changelog

[!IMPORTANT] This project was built by reverse engineering a proprietary binary format. It is not affiliated with or endorsed by Trimble Inc. or SketchUp.

What is OpenSKP?

OpenSKP is the first and only open-source, cross-platform parser for SketchUp binary files — reverse-engineered from both the modern VFF container (SketchUp 2021+) and the classic MFC CArchive container (SketchUp 2013–2020). It gives you full programmatic access to geometry, materials, components, layers, and metadata, with no SketchUp installation and no proprietary SDK required. The same parser and export API also ship as first-class packages for TypeScript, .NET, Dart, and C++ — see the project README for the full cross-language picture.

Features

• Full-fidelity parsing — vertices, edges, faces, normals, UV coordinates, nested component hierarchies, layers/tags, materials, textures, styles, and dynamic-component attributes.
• Both SketchUp file generations — modern VFF (2021+) and legacy MFC (2013–2020) containers, transparently, behind one parse() call.
• Scene baking — an opt-in build_scene() pass resolves the full placed scene graph to world-space, triangulated, export-ready geometry.
• Native multi-format export — glTF (GLB), Wavefront OBJ/MTL, STL, PLY, AutoCAD DXF (3DFACE and Polyface Mesh), IFC4 (BIM/ISO 10303-21 STEP), and JSON — all written from scratch, no third-party CAD/BIM SDK involved. The DXF writer is verified against real desktop AutoCAD, not just lenient DXF readers.
• Streaming / low-memory parsing — peak memory is bounded by the largest single definition, not the whole file.
• Structured observability — opt-in progress reporting and location-carrying parse errors for debugging malformed or unusual files.

Installation

pip install openskp

Or install from source:

git clone https://github.com/iamahsanmehmood/openskp.git
cd openskp/packages/python
pip install -e .

Quick Start

from openskp import SkpFile

# Parse an SKP file
skp = SkpFile.open("model.skp")
model = skp.parse()

# Inspect layers
for layer in model.layers:
    print(f"{layer.name}: rgb({layer.color_r}, {layer.color_g}, {layer.color_b})")

# Inspect definitions (component geometry)
for defn in model.definitions.values():
    print(f"{defn.name}: {len(defn.faces)} faces, {len(defn.vertices)} vertices")

# Inspect resolved, world-space scene hierarchy (opt-in, heavier)
scene = skp.build_scene()
for inst in scene.scene_hierarchy.children:
    print(f"  {inst.name} [{inst.layer}] @ {inst.position_mm}")

Exporting

from openskp.export import glb, obj, stl, ply, dxf, ifc, json_export

# Export to GLB (glTF 2.0 binary) - takes the SkpFile itself
glb.export(skp, "output.glb")

# Export to Wavefront OBJ (+ companion .mtl) - takes a built Scene
obj.export(scene, "output.obj")

# Export to STL (3D Printing ASCII/Binary)
stl.export(scene, "output.stl", binary=True)

# Export to PLY (Stanford 3D Triangle Mesh)
ply.export(scene, "output.ply", binary=True)

# Export to AutoCAD 3D DXF (AutoCAD R2000 compliant, Polyface Mesh by default)
dxf.export(scene, "output.dxf")

# Export to IFC4 / BIM (ISO 10303-21 STEP ASCII format)
ifc.export(scene, "output.ifc")

# Export metadata as JSON - pass scene= to include the resolved hierarchy
meta = json_export.to_dict(model, scene=scene)
json_export.export(model, "output.json", scene=scene)

Package Structure

Requirements

• Python ≥ 3.9
• NumPy ≥ 1.20
• Trimesh ≥ 3.0
• Shapely ≥ 1.8

Used in Production

OpenSKP powers the SketchUp import pipeline for FrameSmart (a 3D collaboration platform with nearly 200 active users) and IngeTrazo (a SketchUp-alternative 3D modeler with a BIM → IFC bridge). Using OpenSKP in your own project? Open an issue or a PR to get added here.

License

MIT — see the root repository for full documentation and multi-language packages.