Running JupyterLab Locally on a MacBook Pro

A simple portable way to run Jupyter notebooks on my local machine with local data.

Enabling rapid prototyping that can be switched over to Google Colab easily by saving and uploading my .ipynb files.

jupyter-lab is a binary that runs a local server that manages different python kernels for isolated UI-driven development.

Install JupyterLab via UV

uv is an awesome package manager for python and it handles virtual environments too. Install docs. It can be thought of as a better pip with venv.

# This installs JupyterLab in an isolated environment and adds jupyter to your PATH (typically ~/.local/bin/).
uv tool install jupyterlab

Launching

Create a notebooks directory and run:

mkdir -p ~/notebooks
cd ~/notebooks && jupyter-lab

JupyterLab auto-opens in your browser at http://localhost:8888.

Custom Kernels for using different Python versions

In notebooks, your kernel uses a predefined python version. There are many situations where you want to run different python versions in different notebooks.

Create a Project Kernel

mkdir -p ~/notebooks/py-3.13
cd ~/notebooks/py-3.13

# Install a specific Python if needed
uv python install 3.13

# Create venv with that Python (to use %pip install (Jupyter's built-in magic) we pass --seed)
uv venv --python 3.13 --seed

# Install ipykernel (uv auto-detects .venv)
uv pip install ipykernel

# Register as a Jupyter kernel
uv run python -m ipykernel install --user \
  --name python3.13 \
  --display-name "Python 3.13"

You can also add common / shared dependencies to this kernel / .venv:

uv pip install --python ~/notebooks/py-3.13/.venv pandas numpy requests

Now “Python 3.13” appears in JupyterLab’s kernel picker.

Managing Secrets with dotenv

Never hardcode API keys in notebooks. Use python-dotenv to load secrets from a .env file:

# In your project venv
%pip install python-dotenv

Create a .env file in your project directory:

# ~/notebooks/my-project/.env
API_KEY=...
DATABASE_URL=postgres://...

Load secrets in your notebook:

from dotenv import load_dotenv
import os

load_dotenv()
api_key = os.getenv("API_KEY")

Cleanup

To remove JupyterLab:

uv tool uninstall jupyterlab

Remove custom kernels:

jupyter kernelspec list
jupyter kernelspec remove <kernel-name>

Wrapping Up

To upgrade, run uv tool upgrade jupyterlab. I’ve found the combination of a local notebook, with a lot of unified memory, to be very powerful.