jupyterlite

Build JupyterLite site

$ jupyter lite build

$ jupyter lite build --contents [notebooks/]

$ jupyter lite serve

$ jupyter lite build --piplite-wheels [package.whl]

$ jupyter lite clean

$ jupyter lite status

$ jupyter lite build --output-dir [dist/]

$ jupyter lite init

jupyter lite command [options]

jupyter lite builds static JupyterLab sites that run entirely in the browser using WebAssembly. No server required—deploy to any static hosting.
build generates HTML, JavaScript, and assets in the output directory. Include notebooks with --contents to pre-populate the file browser. The resulting site runs JupyterLab with Pyodide kernel.
The Pyodide kernel executes Python in WebAssembly. Pre-install packages with --piplite-wheels or install at runtime via piplite. Many scientific Python packages are available.
serve starts a local HTTP server for testing. The build output can deploy to GitHub Pages, Netlify, or any static host. No backend infrastructure needed.
Configuration via jupyter_lite_config.json customizes settings, extensions, and build behavior. init creates template files.

--contents path

Path to content directory for notebooks.

Output directory for built site. Default: _output.

Include pip wheels for piplite.

Custom Pyodide distribution URL.

Port for serve command. Default: 8000.

Source directory for JupyterLite files.

JSON file with settings overrides.

Disable source maps in build.

build

Build JupyterLite static site.

Serve built site locally.

Initialize configuration files.

Show build configuration status.

Remove build artifacts.

Create archive of built site.

List addons and their status.

Not all Python packages work in Pyodide (no native extensions without WASM builds). Performance is slower than native Python. Large notebooks may hit browser memory limits. Some JupyterLab extensions are incompatible.

JupyterLite was released in 2021 by the Project Jupyter team, leveraging Pyodide for browser-based Python. It enables sharing interactive notebooks without server infrastructure. The project grew from experiments with WebAssembly Python interpreters. JupyterLite enables educational content, documentation, and lightweight demos. Development continues alongside JupyterLab improvements.

jupyter-lab(1), pyodide(1), jupyter(1)