K3D-jupyter is a 3D plotting library for Jupyter notebooks. It draws meshes, volumes, voxels, point clouds, vector fields and lines straight from NumPy arrays, in JupyterLab, Colab and VS Code, and exports standalone HTML snapshots that open without a server.
It is used where researchers need to look at 3D data inside the notebook they compute in: medical and biological imaging (CT and microscopy volumes), fluid dynamics, micromagnetics and teaching. Libraries such as ubermag, pyMOR, vedo and sympy_plot_backends use it as a plotting backend.
The project has been developed in the open since 2015 under the MIT license. It was installed from PyPI about 970,000 times in the first ten and a half months of 2026.
Current work: a WebGPU rendering backend replacing WebGL, faster volume rendering for large CT scans, and a physically based path tracer for publication figures. Everything ships in the open-source release; there is no paid edition.
It is used where researchers need to look at 3D data inside the notebook they compute in: medical and biological imaging (CT and microscopy volumes), fluid dynamics, micromagnetics and teaching. Libraries such as ubermag, pyMOR, vedo and sympy_plot_backends use it as a plotting backend.
The project has been developed in the open since 2015 under the MIT license. It was installed from PyPI about 970,000 times in the first ten and a half months of 2026.
Current work: a WebGPU rendering backend replacing WebGL, faster volume rendering for large CT scans, and a physically based path tracer for publication figures. Everything ships in the open-source release; there is no paid edition.