Awesome CAD โ Curated List of Open-Source CAD Projects
A curated list of valuable open-source CAD (Computer-Aided Design) software, libraries, and tools, organized by category.
This list focuses on practical, reusable, and technically interesting CAD projects for developers, engineers, and researchers.
Inspired by the awesome-list philosophy.
๐ Table of Contents
โ๏ธ 2D CAD Software
Desktop 2D CAD
LibreCAD
https://github.com/LibreCAD/LibreCAD
A mature, free 2D CAD application focused on DXF drafting.
Key Features
- Native DXF support
- Layers, blocks, snapping, dimensions
- Cross-platform (Qt)
- AutoCAD LTโstyle workflow
QCAD Community Edition
Professional-grade 2D drafting software with a long development history.
Key Features
- Excellent DXF compatibility
- ECMAScript scripting
- Clean UI and stable core
- Widely used in industry
CADCraft
https://github.com/storytold/cadcraft
An open-source clean-room reimplementation of Autodesk AutoCAD in pure Rust, bringing the classic command-line drafting workflow to modern cross-platform native apps.
Key Features
- AutoCAD-style command line, object snaps, layers, dimensions, hatches, and blocks
- Native DXF (R12โ2018, ASCII & binary) and DWG read/write
- Parametric geometric and dimensional constraints with conflict detection
- Paper-space layouts, viewports, and PDF / SVG / PNG export
- Cross-platform desktop (macOS, Windows, Linux, FreeBSD) and web (WebAssembly)
- Agent-friendly: MCP server, JSON control channel, and CLI for automation
Web-Based 2D CAD
cad-viewer
https://github.com/mlightcad/cad-viewer
A modern browser-native CAD viewer that parses DXF/DWG directly on the client.
Key Features
- Client-side DXF/DWG parsing (no upload)
- Optimized for very large drawings
- Accurate snapping & layers
- WebGL / Three.js based
- Privacy-first architecture
Maker.js
https://github.com/microsoft/maker.js
A JavaScript library for programmatic 2D CAD geometry.
Key Features
- Declarative geometry modeling
- DXF / SVG output
- Perfect for CNC & laser cutting
- Strong TypeScript support
๐งฉ 3D CAD Modelers
Desktop 3D CAD
FreeCAD
https://github.com/FreeCAD/FreeCAD
A full-featured parametric 3D CAD system for engineering and design.
Key Features
- B-Rep modeling (OpenCascade)
- Sketch constraints & parametrics
- Assemblies, FEM, CAM, TechDraw
- Massive plugin ecosystem
Blender
https://github.com/blender/blender
A world-class mesh-based 3D creation suite (not B-Rep CAD).
Key Features
- Advanced mesh modeling
- Geometry Nodes (procedural)
- Python scripting
- Industry standard for mesh workflows
OpenSCAD
https://github.com/openscad/openscad
A script-based solid modeling tool emphasizing precision and reproducibility.
Key Features
- Text-based parametric modeling
- Exact geometry
- Excellent for 3D printing
- Deterministic CAD output
build123d
https://github.com/gumyr/build123d
A modern Pythonic parametric CAD framework built on OpenCascade.
Key Features
- Python-first CAD API
- Clean, composable design
- B-Rep modeling
- Ideal for automation & pipelines
cadquery-editor
https://github.com/CadQuery/CQ-editor
A desktop IDE for script-driven parametric 3D CAD based on CadQuery and OpenCascade.
Key Features
- Script-driven parametric modeling
- Python-based CAD logic
- Built on CadQuery + OpenCascade
- Interactive 3D preview
- Popular in engineering automation and design pipelines
SolveSpace
https://github.com/solvespace/solvespace
A lightweight constraint-based parametric 2D/3D CAD system focused on precision and simplicity.
Key Features
- Strong geometric constraint solver (core strength)
- 2D sketching and 3D solid modeling
- Parametric dimensions and relations
- Very small, fast, and deterministic
- Popular for mechanical sketches, linkages, and educational use
- Fully open-source and self-contained
OpenCAD Studio
https://github.com/HakanSeven12/OpenCADStudio
An open-source 2D drafting and 3D modeling application built in Rust, with native DWG/DXF support across desktop and web.
Key Features
- Native DWG/DXF read/write (R14 through 2018)
- 2D drafting: lines, polylines, splines, hatches, snaps, layers, blocks, dimensions
- Kernel-backed 3D modeling: extrude, revolve, sweep, loft, Boolean operations
- GPU-accelerated rendering via wgpu
- Cross-platform desktop (Windows, Linux, macOS) and web (WebAssembly)
- Plugin system, headless automation, and MCP endpoint for AI agents
Web-Based 3D CAD
jsketcher
https://github.com/xibyte/jsketcher
A parametric, sketch-based CAD system running in the browser.
Key Features
- Constraint-driven sketches
- OpenCascade via WebAssembly
- Browser-native CAD kernel
- One of the most advanced web CAD projects
chili3d
https://github.com/xiangechen/chili3d
A lightweight web-based 2D/3D CAD editor focused on sketches and parametric modeling.
Key Features
- Sketch-driven CAD workflow
- 2D + 3D modeling
- Runs fully in the browser
- Three.js rendering
- Ideal for CAD experimentation
๐งโ๐ป Code-Driven / Scriptable CAD
Design CAD models using code instead of GUI interactions โ ideal for reproducible, parametric, automated, and version-controlled workflows.
CadQuery
https://github.com/CadQuery/cadquery
A Python library for parametric, script-based CAD modeling built on OpenCascade.
Key Features
- Pythonic CAD API
- B-Rep modeling via OpenCascade
- Constraint-friendly workflows
- Widely used in engineering automation
- Strong ecosystem (CQ-editor, exporters)
OpenJSCAD
https://github.com/jscad/OpenJSCAD.org
A JavaScript-based CAD system for scriptable solid modeling in the browser and Node.js.
Key Features
- JavaScript CAD modeling
- Browser and CLI support
- CSG-based modeling
- STL / AMF / 3MF export
- Popular in maker and web CAD communities
replicad
https://github.com/sgenoud/replicad
A JavaScript library for browser-native, code-driven 3D CAD modeling built on OpenCascade.
Key Features
- OpenCascade abstraction โ full B-Rep modeling power in the browser
- CadQuery-inspired streamlined API: sketch โ shape โ modify workflow
- Browser-native execution with no server dependency
- STEP export, fillets, chamfers, and other industrial-grade kernel features
- Modern JavaScript ecosystem, integrates with existing web tooling
Manifold
https://github.com/azrecol/manifold
A fast and robust mesh geometry library supporting CSG and implicit (SDF) operations.
Key Features
- High-performance mesh booleans
- CSG and SDF-based modeling
- Robust handling of non-manifold geometry
- Used internally by OpenSCAD
- Suitable for CAD โ mesh pipelines
SolidPython
https://github.com/SolidCode/SolidPython
A Python frontend that generates OpenSCAD-compatible code.
Key Features
- Pythonic wrapper for OpenSCAD
- Parametric & reproducible designs
- Easy integration with Python tooling
- Ideal for quick scripting and automation
๐ค AI-Assisted Design
CADAM
https://github.com/Adam-CAD/CADAM
An open-source AI-powered text-to-CAD web application that turns natural language and images into parametric 3D models in the browser.
Key Features
- Natural language input to describe 3D models
- Image-assisted model generation
- Parametric controls with interactive sliders
- Real-time 3D preview in browser
- Export to .STL and .SCAD formats
- Browser-based with no installation required
vcad
Open-source parametric CAD for the AI era โ a full B-Rep modeler with web, desktop, and CLI surfaces, designed for human and agent workflows alike.
Key Features
@vcad/mcpserver:create_cad_document,export_cad,inspect_cad,gym_step,gym_reset- Rust/WASM B-Rep kernel: primitives, booleans, fillets, chamfers, shell, sketch constraints
- Extrude, revolve, sweep, loft; assembly with joints and forward kinematics
- Simulation (phyz) with gym-style RL interface
- STEP import; STL, GLB, STEP, DXF export; direct B-Rep ray tracing
- Web app at vcad.io, desktop installers, and
vcad-cli
text-to-cad (CAD Skills)
https://github.com/earthtojake/text-to-cad
A library of agent skills for CAD, robotics, and hardware design โ generate, inspect, source, slice, and hand off CAD and robot-description artifacts from local project files.
Key Features
- CAD skill: plain-language or image-driven modeling with STEP, STL, 3MF, and GLB export (build123d / OpenCascade)
- CAD Viewer: local browser previews for CAD, G-code, and robot files
- Robot description: URDF, SRDF (MoveIt), and SDF skills
- Fabrication: G-code slicing, SendCutSend checks, Bambu Labs print workflows
- Off-the-shelf parts via step.parts
- Install via Skills CLI, Codex, or Claude Code plugins
NURBGen
https://github.com/SadilKhan/NURBGen
The first framework to generate industry-standard NURBS surfaces directly from text prompts, producing editable, parametric CAD models convertible to STEP format (AAAI 2026).
Key Features
- LLM-driven text-to-CAD via fine-tuned Qwen3-4B (LoRA adapter on Hugging Face)
- Direct NURBS surface generation for high-fidelity, industry-standard geometry
- CLI and Python API inference (ms-swift, HuggingFace / PEFT)
- Export to STEP and STL from generated NURBS representations
- Evaluation pipeline for comparing predicted and ground-truth STEP models
Forgent3D
https://github.com/forgent3d/forgent3d
A local AI CAD companion for building, previewing, and iterating parametric 3D models with coding agents.
Key Features
- Parametric models driven by
part.pyorasm.xmlplusparams.json - Live local preview and rebuild loop in a Three.js desktop viewer
- AI-agent friendly with built-in project skills and MCP tooling
- Geometry-first validation via MJCF, with screenshots and bounding-box feedback
- Multi-body assemblies with joints, constraints, and optional MuJoCo simulation
- Bundled desktop app with embedded build123d CAD runtime
๐ Web-Based CAD / DXF Viewers
cad-viewer
https://github.com/mlightcad/cad-viewer
High-performance CAD viewer focused on correctness and scale.
Key Features
- DXF/DWG support
- Large drawing optimization
- Layer control & snapping
- No server dependency
dxf-viewer
https://github.com/vagran/dxf-viewer
A minimal DXF viewer for quick embedding.
Key Features
- Lightweight
- Canvas-based rendering
- Simple DXF support
three-dxf
https://github.com/gdsestimating/three-dxf
DXF loader for Three.js scenes.
Key Features
- DXF โ Three.js conversion
- WebGL-friendly
- Easy integration with 3D viewers
gemini-viewer-examples
https://github.com/pattern-x/gemini-viewer-examples
Example projects and demos showcasing the Gemini CAD viewer, a WebGL-based 3D viewer optimized for CAD data.
Key Features
- Interactive 3D model viewing in the browser
- WebGL/WebGPU rendering
- Includes examples for various CAD formats
- Good starting point for integrating web CAD viewers
๐ DXF / DWG Parsers & Libraries
LibreDWG
https://github.com/LibreDWG/libredwg
A C library for reading and writing DWG files, part of the GNU project.
Key Features
- Native DWG read/write support
- Covers multiple AutoCAD DWG versions
- Low-level, spec-oriented implementation
- Suitable for CAD import/export pipelines
ezdxf
https://github.com/mozman/ezdxf
A Python library for reading, writing, and modifying DXF files.
Key Features
- High-level Python API
- Excellent DXF spec coverage
- Actively maintained
- Ideal for CAD automation and data processing
- Popular in scripting and batch workflows
ACadSharp
https://github.com/DomCR/ACadSharp
A .NET library for reading, writing, and manipulating DXF and DWG files without requiring AutoCAD.
Key Features
- Native C# / .NET API
- Read and write DXF (ASCII & binary) and DWG files
- Strong support for CAD structures (Entities, Blocks, Layers, Tables, Styles)
- Multi-version DXF/DWG compatibility
- Fine-grained control over geometry and document data
- No AutoCAD dependency
- Open-source (MIT license), suitable for automation and CAD tooling
dxflib
https://github.com/LibreCAD/dxflib
A widely used C++ DXF parsing library.
Key Features
- Accurate DXF spec support
- Mature & battle-tested
- Used by LibreCAD
libdxfrw
https://github.com/LibreCAD/libdxfrw
An open-source DXF/DWG read/write library.
Key Features
- DXF + DWG support
- Multiple AutoCAD versions
- Core dependency for LibreCAD
dxf-parser
https://github.com/gdsestimating/dxf-parser
A JavaScript DXF parser for web and Node.js.
Key Features
- Browser-friendly
- Common in web CAD viewers
- Easy to extend
dxf-json
https://github.com/dotoritos-kim/dxf-json
A tool for converting DXF files to JSON format for easier processing in web and scripting environments.
Key Features
- Converts DXF to structured JSON
- Browser and Node.js friendly
- Lightweight and easy to integrate
๐ CAD File Converters & Interoperability
Tools focused on converting and exchanging data between CAD file formats โ especially translating proprietary native formats (SolidWorks, CATIA, NX, Creo, Inventor) into neutral interchange formats like STEP.
cadmpeg
https://github.com/cadmpeg/cadmpeg
A command-line tool for inspecting and converting native CAD files, built in Rust with a clean-room approach โ an open-source alternative to proprietary vendor translation SDKs.
Key Features
- Broad native format support: SolidWorks
.sldprt, Inventor.ipt/.iam, Fusion 360.f3d, CATIA V5.CATPart, NX / Creo.prt, Rhino.3dm, ASM/ACIS.sat/.smt, FreeCAD.FCStd - Convert native CAD files to STEP (and inspect, measure, and diff geometry between files)
- Clean-room decoders โ format knowledge from public documentation and legally possessed files, no vendor SDKs
- Per-format support-level ladder with documented byte-level format profiles
- Rust workspace with reusable crates for embedding native CAD import into applications
- Apache 2.0 (code), CC BY 4.0 (docs), cross-platform install (Homebrew, installer scripts, cargo)
๐งฐ CAD Libraries & Kernels
Geometry & Modeling Kernels
Open Cascade Technology (OCCT)
https://github.com/Open-Cascade-SAS/occt
An industrial-grade B-Rep geometry kernel.
Key Features
-
Solid & surface modeling: robust B-Rep modeling with Booleans, fillets, chamfers, offsets, sweeps, drafts, and advanced construction tools.
-
Rich geometric algorithms: point/surface projection, curve/surface intersections, analytical and approximate constructions.
-
NURBS & parametric geometry: support for NURBS curves and surfaces, transformations, and analytic primitives (planes, cylinders, cones, spheres, etc.).
-
Data interoperability: import/export for STEP, IGES, STL, VRML, BREP plus extended exchange with attributes and assembly (XCAF).
-
Shape healing & model repair: tools to correct inconsistencies in imported CAD models.
-
Visualization & rendering: interactive 3D display, selection, shading, real-time ray tracing, layered display controls.
-
Mesh handling: conversion between exact geometry and tessellated meshes for display and rapid prototyping workflows.
-
Application framework: document management (OCAF), undo/redo, attribute tracking, and extensible architecture.
-
Modular C++ class libraries designed for rapid development of domain-specific CAD/CAM/CAE applications.
pythonOCC
https://github.com/tpaviot/pythonocc
Python bindings for OpenCascade.
Key Features
- Rapid CAD prototyping
- Scriptable B-Rep modeling
- Research & automation friendly
TiGL
https://github.com/DLR-SC/tigl
A parametric geometry kernel for aircraft design based on CPACS and OpenCASCADE.
Key Features
- Parametric aircraft geometry (wings, fuselages, control surfaces)
- CPACS-driven modeling workflow
- NURBS-based surface generation (built on OCCT)
- Geometry queries (surface points, intersections, sections)
- Export to STEP, IGES, STL, VTK, Collada
- Multi-language bindings (C, C++, Python, Java, MATLAB)
- Includes TiGL Viewer for geometry visualization
libfive
https://github.com/libfive/libfive
A functional CAD kernel based on implicit geometry (SDF).
Key Features
- Infinite resolution geometry
- Functional modeling paradigm
- Excellent for generative CAD
USD BRep
https://github.com/NVIDIA-Omniverse/usd-brep
A standalone NURBS-based boundary representation (B-Rep) solid modeling library from NVIDIA Omniverse, with C++ and Python APIs and native OpenUSD integration.
Key Features
- NURBS-based B-Rep solid modeling kernel
- C++ and Python APIs, including Boolean operations
- OpenUSD schema plugin (omniSolid) for precise B-Rep geometry in USD scenes
- SMLib-to-USD mesh authoring workflow and B-Rep asset healing tools
- Interactive Python GUI for exploration
- Cross-platform (Windows, Linux, macOS)
Constraint Solvers
KittyCAD / ezpz
https://github.com/KittyCAD/ezpz
A modern geometric constraint solver for parametric CAD.
Key Features
- Rust + WebAssembly
- Designed for sketches
- Web-CAD ready
NeoGeoSolver.NET
https://github.com/NoteCAD/NeoGeoSolver.NET
A geometric constraint solver written in C#.
Key Features
- Sketch constraint solving
- Educational & practical
- Clean solver design
Frontier
https://github.com/Geoplexity/Frontier
A research-oriented geometric constraint solving framework for parametric CAD systems.
Key Features
- Focused on geometric constraint solving
- Designed for sketch-based CAD
- Emphasizes robustness and correctness
- Suitable for academic and experimental CAD engines
- Clear separation between geometry and solver logic
pygeosolve
https://github.com/SeanDS/pygeosolve
A Python-based geometric constraint solver using numerical optimization techniques.
Key Features
- Written in Python
- Uses numerical methods to solve constraints
- Suitable for prototyping and research
- Easy to integrate with Python CAD workflows
- Helpful for understanding constraint-solving fundamentals
Mesh Processing & Boolean Engines
CGAL
The gold standard for computational geometry.
Key Features
- Robust mesh booleans
- Exact arithmetic
- Research & industrial use
libigl
https://github.com/libigl/libigl
A simple geometry processing library.
Key Features
- Mesh analysis & deformation
- Academic-friendly
- Easy integration
Manifold
https://github.com/elalish/manifold
A fast and robust mesh boolean engine.
Key Features
- Handles non-manifold input
- Used by OpenSCAD
- Performance-focused
OpenVDB
https://github.com/AcademySoftwareFoundation/openvdb
A sparse volumetric representation library.
Key Features
- Voxel-based booleans
- CAD โ VFX pipelines
- Extremely scalable
๐ Specialized CAD Tools
OpenVSP
https://github.com/OpenVSP/OpenVSP
A parametric aircraft geometry modeling tool developed by NASA.
Key Features
- Aircraft-specific parametric modeling
- Aerospace-grade geometry
- Industry & research usage
Fritzing
https://github.com/fritzing/fritzing-app
An electronics CAD and prototyping tool.
Key Features
- Schematic + PCB + breadboard views
- Maker-friendly
- Open hardware ecosystem
๐ CAD Datasets
CAD datasets organized by representation and supervision type, covering mesh-based geometry, parametric / B-Rep CAD, and language-guided (Text-to-CAD) datasets.
๐บ Mesh-Based CAD / Shape Datasets
Datasets where geometry is represented primarily as triangle meshes or point clouds.
Commonly used in 3D vision, shape generation, and geometry learning, but not ideal for parametric CAD reconstruction.
ShapeNet (CAD subset)
https://github.com/ShapeNet/ShapeNet
A massive dataset of 3D shapes, including many CAD-like objects.
Key Features
- Large-scale mesh dataset
- Category-level annotations
- Widely used ML benchmark
- Not parametric or editable CAD
PartNet
https://github.com/daerduoCarey/PartNet
A structured dataset with hierarchical part annotations built on ShapeNet.
Key Features
- Fine-grained part decomposition
- Semantic part labels
- Assembly-aware structure
- Mesh-based representation
Text2Shape
https://github.com/kchen92/text2shape
A dataset pairing natural language descriptions with 3D shapes.
Key Features
- Text โ mesh pairs
- Early benchmark for text-driven 3D generation
- Useful for multimodal pretraining
- Not parametric CAD
๐ Parametric / B-Rep CAD Datasets
Datasets that represent geometry as procedural CAD programs, feature trees, or B-Rep solids.
These are critical for editable CAD, design automation, and engineering-focused ML.
ABC Dataset (Autodesk)
https://github.com/AutodeskAILab/abc
A large-scale dataset of real-world CAD models.
Key Features
- B-Rep + mesh representations
- Industry-grade CAD geometry
- High-quality solids
- Popular benchmark for CAD reconstruction
Fusion 360 Gallery Dataset
https://github.com/AutodeskAILab/Fusion360GalleryDataset
A dataset of parametric CAD models with full design history.
Key Features
- Feature trees and construction sequences
- Sketches + operations
- Ideal for learning parametric modeling
- Strong foundation for CAD ML research
DeepCAD
https://github.com/ChrisWu1997/DeepCAD
A parametric CAD dataset with explicit procedural construction sequences.
Key Features
- Programmatic CAD modeling steps
- Sketch-based operations
- Clean parametric representation
- Ideal for sequence learning and CAD synthesis
WHUCAD
https://github.com/fazhihe/WHUCAD
Paper: https://journals.sagepub.com/doi/10.3233/ICA-240744
A large-scale feature-level parametric CAD dataset designed to model advanced CAD operations beyond simple sketchโextrude workflows.
Key Features
- Parametric CAD programs with structured command sequences
- Includes advanced modeling features
- Bridges low-level commands and high-level feature representations
- Suitable for procedural CAD generation and feature learning
Zero-To-CAD 1M
https://huggingface.co/datasets/ADSKAILab/Zero-To-CAD-1m
Paper: https://arxiv.org/abs/2604.24479
A million-scale dataset of executable CadQuery construction sequences synthesized entirely without real-world CAD data, produced by an agentic LLM in a feedback-driven CadQuery environment.
Key Features
- ~1M geometrically validated samples (train / validation / test splits)
- Executable, interpretable CadQuery Python per sample with operation JSON metadata
- Multi-view renders (8ร256ร256), plus STL and STEP exports per model
- Broad operation coverage: booleans, fillets, chamfers, shells, lofts, sweeps, patterns, and more
- 65 synthetic part categories; precomputed DINOv3 embeddings and FAISS nearest-neighbor index
- Apache 2.0 license (Autodesk AI Lab)
Primary Use Cases
- Training CAD sequence and program-synthesis models
- Image-to-CAD reconstruction from multi-view renders
- Benchmarking generative CAD models on diverse, validated geometry
๐ Text-Guided / Text-to-CAD Datasets
Datasets that align natural language with CAD models or CAD programs, enabling research in
Text โ CAD generation, language-guided modeling, and multimodal CAD understanding.
Text2CAD Dataset
https://sadilkhan.github.io/text2cad-project/
A large-scale dataset explicitly designed for Text-to-CAD research, aligning natural language with parametric CAD models.
Key Features
- Natural language โ parametric CAD pairs
- CAD represented as procedural programs
- Fine-grained design intent annotations
- Supports Text โ CAD and CAD โ Text tasks
Primary Use Cases
- Text-to-parametric CAD generation
- Language-driven CAD editing
- CAD retrieval from natural language
- Multimodal representation learning
Omni-CAD Dataset
https://huggingface.co/datasets/jingwei-xu-00/Omni-CAD
A large-scale multimodal CAD dataset designed for unified learning across geometry, CAD programs, and natural language.
Key Features
- Multimodal CAD representation (geometry + program + language)
- Supports Text โ CAD โ Geometry tasks
- Includes parametric CAD construction sequences
- Designed for training foundation models for CAD
- Suitable for CAD generation, understanding, and editing
Primary Use Cases
- Text-to-CAD generation
- CAD program synthesis
- Multimodal CAD representation learning
- CAD understanding with language supervision
ShapeTalk
https://github.com/ShapeTalk/ShapeTalk
A dataset of language grounded in 3D geometry.
Key Features
- Referring expressions for shapes
- Language grounding in geometry
- Useful for semantic understanding
- Often combined with CAD datasets
Fusion 360 Gallery + Language (Derived)
https://github.com/AutodeskAILab/Fusion360GalleryDataset
While not natively textual, Fusion 360 Gallery is frequently augmented with synthetic or human-written descriptions for Text-to-CAD experiments.
Key Features
- Parametric CAD + feature trees
- Common base dataset for text annotation
- Used in weakly supervised Text-to-CAD research
- Bridges classic CAD ML and language models
๐ค Contributing
Contributions are welcome!