Installation

You can install metakernel through pip:

pip install metakernel --upgrade

Installing a kernel

MetaKernel-based kernels expose an install subcommand that registers the kernel with Jupyter:

python -m my_kernel install --user

All flags are forwarded to jupyter kernelspec install. Common options:

--user

Install into the current user’s Jupyter data directory.

--sys-prefix

Install into the active Python environment’s prefix (useful with conda/venv).

--prefix <path>

Install into an arbitrary prefix.

--name <name>

Override the kernel directory name used by Jupyter.

--display-name <name>

Set the display name shown in the JupyterLab kernel picker. This is written into the display_name field of the installed kernel.json and is useful when running multiple installations of the same kernel side-by-side (for example, one per environment):

python -m my_kernel install --user --name my-kernel-dev --display-name "My Kernel (dev)"