Introducing the JupyterLab Visual Debugger and How to Install It

Jupyter notebooks offer great interactivity but lack built‑in visual debugging, which is essential for large codebases; the Jupyter team therefore released the first version of a visual debugger for JupyterLab.

The new debugger allows setting breakpoints, viewing variables, and navigating call stacks directly within the notebook interface.

To use it, install the front‑end extension with:

<code>jupyter labextension install @jupyterlab/debugger</code>

For the back‑end, the kernel must support the Jupyter Debug Protocol; currently only xeus-python can be used:

<code>conda install xeus-python -c conda-forge</code>

After installing both components, the debugger works out‑of‑the‑box, and an online demo is available without any local installation.

Xeus‑python is the first kernel to support Jupyter debugging. It is a C++ implementation of the Jupyter kernel protocol that enables building language kernels with plug‑in concurrency models and a lightweight codebase, making it easy to iterate and update.

The JupyterLab debugger extension provides IDE‑like features such as a side‑panel file explorer, source preview, breakpoint setting next to code cells, and visual markers indicating the current execution point.

It also supports viewing previously executed cells in read‑only mode, debugging multiple notebooks simultaneously, and inspecting variables via tree and table viewers.

Future plans include richer variable rendering and conditional breakpoints.

The article also mentions a VS Code visual debugging tool that visualizes data structures (e.g., linked lists, trees, plots) during execution, offering customizable visualizers like Plotly, Tree, and Grid, though Python support is still under discussion.

Overall, the JupyterLab visual debugger and related tools aim to bring IDE‑level debugging capabilities to the interactive notebook environment.