JupyterLab 3.0: New Features, Installation Options, and Extension Improvements

JupyterLab 3.0, the next‑generation Jupyter Notebook interface, has been released with many new features and substantial improvements to its extension system.

Key improvements include a visual debugger, support for multiple display languages, a notebook directory view, and an enhanced extension system.

Installation methods

JupyterLab 3.0 can be installed in three ways:

Using pip: pip install jupyterlab==3

Using mamba: mamba install -c conda-forge jupyterlab=3

Using conda: conda install -c conda-forge jupyterlab=3

Users should verify third‑party extensions for compatibility and update them if necessary.

New features

Visual debugger – JupyterLab now includes a visual debugger that works with kernels supporting debugging, such as the Xeus‑Python kernel.

Directory extension – A new directory panel makes it easier to browse document structure.

Multilingual UI – Users can install language packs (e.g., Chinese) via pip install jupyterlab-language-pack-zh-CN to change the interface language.

Simple interaction mode improvements – The classic single‑document mode is smoother and can be toggled via the status bar, view menu, command palette, or shortcut Ctrl/Cmd+Shift+D.

Mobile device support – The layout becomes more compact on small screens, automatically switching to simple interaction mode when the window is reduced.

Pre‑built extensions – Extensions are now distributed as pre‑built Python packages on PyPI or conda‑forge, allowing installation with pip, conda, or mamba without rebuilding JupyterLab or installing Node.js. Example: installing ipywidgets 7.6.0 via pip or conda enables it automatically in both classic Notebook and JupyterLab 3.0.

Extension author workflow – The TypeScript extension cookiecutter has been updated to scaffold pre‑built extensions with all necessary tooling, simplifying development.

For a complete list of changes, including bug fixes, refer to the detailed changelog linked in the original article.