jupyter-live-kernel

Use a live Jupyter kernel via the jupyter_live_kernel.py script's execute command, which keeps variables, imports, and objects alive across calls. Point it at a notebook path on a running JupyterLab server and each execution shares the same kernel session.

Use a live Jupyter kernel for iterative exploration, data science, and tasks where state must persist between steps. Use one-shot execution for stateless scripts that need no memory of prior runs, such as single file operations or web searches.

Run the variables subcommand with the notebook path to list all live variables, or use variables preview with a specific name to inspect one value. Both return structured JSON output, and the --compact flag reduces token usage.

The kernel needs time to initialize after the server starts, so the first execution often exceeds the default 30-second timeout. Retry the command or pass a larger timeout such as --timeout 120 for initial setup or heavy computation.

The kernel uses JupyterLab's Python environment, so packages must be installed into that environment before use. Install JupyterLab itself with uv tool install jupyterlab, then add any additional packages to the same tool environment.