jupyter-lab

Start JupyterLab

$ jupyter lab

$ jupyter lab [path/to/notebook].ipynb

$ jupyter lab --notebook-dir [path/to/directory]

$ jupyter lab --debug

jupyter-lab [OPTIONS] [PATH]

--port=<port>
    Port for server to listen on (default: 8888)

--ip=<ip>
    IP address to listen on (default: localhost)

--no-browser
    Do not open browser after starting server

--notebook-dir=<dir>
    Working directory for notebooks (default: current dir)

--generate-config
    Generate default config file ~/.jupyter/jupyter_lab_config.py

--config-dir=<dir>
    Directory for config files (default: ~/.jupyter)

--allow-root
    Allow running as root (dangerous)

--LabApp.token=<token>
    Token for authentication

--dev-mode
    Enable developer mode for extensions

--ServerApp.open_browser=<bool>
    Open browser automatically

JupyterLab is a flexible, extensible, open-source web-based interactive development environment for working with Jupyter notebooks, code, and data. It provides a notebook-centric interface with support for multiple documents in a tabbed layout, including notebooks, text editors, terminals, and custom components. Unlike the classic Jupyter Notebook, JupyterLab offers a unified interface that scales from individual notebooks to full-fledged computing environments.

Key features include a file browser, console, variable inspector, rich text editing, image viewing, and extension support for languages like Python, R, Julia, and more. It supports kernels for over 100 programming languages and integrates with tools like Git, VS Code extensions, and big data frameworks. Launched via the command line, it starts a local Tornado-based web server, accessible via browser, ideal for data science, machine learning, and scientific computing workflows.

JupyterLab unifies and generalizes the classic notebook experience, enabling reproducible research and collaborative development.

Requires Python 3.6+ and JupyterLab installed via pip or conda. Running on public IP exposes notebooks publicly—use tokens or passwords. Not for production; consider JupyterHub for multi-user setups. Extensions may require rebuild with jupyter lab build.

Install via jupyter labextension install or pip; rebuild with jupyter lab build.

Supports build, clean, extension subcommands for management.

JupyterLab originated from Project Jupyter, first alpha released in 2016, stable 1.0 in 2019. Evolved from Jupyter Notebook (2011) to provide a more IDE-like experience. Maintained by Jupyter team with contributions from NumFOCUS; latest versions focus on LSP integration and performance.

jupyter(1), jupyter-notebook(1), jupyterhub(1), python(1)