Jupyter for CadQuery / build123d

The Python package provides a Jupyterlab extension and a JupyterServer extension to view CadQuery, build123d, and OCP objects in JupyterLab

Current version: v5.1.0 (2026-09-11)

The full documentation lives at bernhard-42.github.io/ocp_viewer_docs — the Jupyter CadQuery chapter covers this viewer's specifics; this README gets you installed and running, and links into the rest.

Click on the "launch binder" icon to start Jupyter-CadQuery on binder:

(Due to security restrictions, the measurement feature does not work on binder)

Installation

• CadQuery

  1. Create and activate a virtual conda environment

     mamba create -n jcq python=3.12
     mamba activate jcq
     

  2. Install latest cadquery master

     mamba install -c conda-forge -c cadquery cadquery=master
     

  3. Install Jupyter CadQuery

     pip install jupyter-cadquery
     

• build123d

  1. Create and activate a virtual environment (conda, pyenv-virtualenv, uv, ...)

  2. Install build123d

     pip install build123d
     

  3. Install Jupyter CadQuery

     pip install jupyter-cadquery
     

Jupyter CadQuery requires JupyterLab 4 (jupyterlab>=4.6.2,<5) and installs it if it is missing. The JupyterLab extension is prebuilt and ships inside the cad-viewer-widget dependency: no jupyter labextension step, no Node.js.

Known issue: ipykernel 7.x can stall a running notebook (a cell stays at [*] with an idle kernel). Until the ipykernel release that carries the fix, install pip install "ipykernel<7" — details and the workaround in Troubleshooting.

Verify the installation

jupyter lab extension list        # must list:  jupyter_cadquery 5.1.0 OK   (the viewer frontend)
jupyter server extension list     # must list:  jupyter_cadquery 5.1.0 OK   (the measurement backend)

A line "Extension package jupyter_cadquery took N s to import" is OCP and VTK loading, not an error. On a Mac the very first run can take minutes for the same reason.

Changed behavior from 4.x

• The complete sidecar space will now be used for the viewer. For the old behaviour set cad_width and height
• The default for show is reset_camera=Camera.KEEP now (unless set in ~/.jcq_config differently).
  • The viewer warns if the new object is too large or too small. To disable the warnings, use ignore_camera_warnings()
  • The old behavior can be achieved with set_defaults(reset_camera=Camera.RESET)
• from jupyter_cadquery import * still works. For portable code with other viewers from the ocp viewer ecosystem use from ocp_viewer_core.viewer import *
• mate_scale does not work any more, use helper_scale
• The viewer now remembers paths in the tree that are shown/hidden and will show/hide the same paths for the newly shown CAD object

For all new features and fixes see CHANGELOG

First run

jupyter lab

and in a notebook:

from build123d import *
from jupyter_cadquery import *

open_viewer("CAD")
show(Box(1, 2, 3))

open_viewer puts the viewer into a sidecar panel next to the notebook; without it, show draws into the cell. From here on, everything is in the documentation.

Documentation

The full documentation lives at bernhard-42.github.io/ocp_viewer_docs — the Jupyter CadQuery chapter covers this viewer's specifics; everything below is a deep link into it.

Getting started

• Overview — what it does, in pictures
• Installation
• Sidecars, windows and cells — where a viewer lives, its size, addressing one of several, managing viewers
• Best practices for configuring

Working in the notebook

• Working in the notebook — auto display, reading a pick back into Python, what the notebook remembers
• Replay — step through the calls that built a CadQuery object
• Export — a viewer as a standalone HTML page, a notebook as HTML, PNG
• Workspace Config — the settings in ~/.jcq_config

Working with the viewer

• The CAD Viewer window and mouse and keys
• Measurement tools
• Object selection tool
• Physically based rendering Studio
• ImageFace — use a 2-D image as a reference plane

Python show* commands

• Use the show command
• Use the show_object command
• Use the push_object and show_objects commands
• Use the show_all command
• Use the set_viewer_config command

Python API reference

• Jupyter CadQuery API — open_viewer, close_viewer, export_html, replay, get_pick, …
• Additional Python API (show_clear, save_screenshot, status, …)
• Animation
• Color maps
• Enums reference (Camera, Collapse, Render, AnalysisTool, UiTab, Studio*)

Help

• Troubleshooting
• How the pieces fit together

Demo Notebooks

Standard examples

• A run through of many features
• Standard CadQuery examples in Jupyter CadQuery
• An OCP example (the OCC bottle)
• CadQuery Sketch support
• build123d examples

Animated examples (requires pip install cadquery-massembly matplotlib):

• Rotating disk arm
• Hexapod
• Jansen Linkage
• CadQuery's door assembly example
• A nested Assembly

Changes

see CHANGELOG.md

Licence

Apache-2.0.