Deploy your first JupyterLite website on GitHub Pages#


If you first want to get familiar with the interface, check out the User Guide.

JupyterLite can easily be deployed on GitHub Pages, using the jupyter-lite CLI to add content and extensions.


Deploying to GitHub Pages requires a Github account.

Generate a new repository from the template#

The jupyterlite demo repository is a template to easily:

  • build a JupyterLite website using prebuilt JupyterLite assets bundling a collection of pre-existing Jupyter Notebooks as part of the distribution

  • deploy the website to GitHub Pages

The process is automated using Github Actions.

Click on “Use this template” to generate a repository of your own from this template:

an animated gif to show how to use the provide repo template to generate a new website

From the Actions tab on your repository, ensure that workflows are enabled. When you make a commit to the main branch, a Github Action will build your JupyterLite release and deploy it to the repository’s Github Pages. By default, the Github Pages site will be found at You can also check the URL from the Repository Settings tab Pages menu item.

If the deployment failed, go to “Settings - Actions - General”, in the “Workflow permissions” section, check “Read and write permissions”. Check that you have Github Pages enabled for your repository: from your repository Settings tab, select the Pages menu item and ensure that the source is set to GitHub Actions:

a screenshot showing the GitHub Actions configuration option for deploying to GitHub Pages

When you commit a file, an updated website will be built and published on Github Pages.


Note that it may take a few minutes for the Github Pages site to be updated. Do a hard refresh on your Github Pages site in your web browser to see the new version of the website.


Alternatively, you can use the JupyterLite demo using xeus-python to publish a deployment on Github pages that uses xeus-python by default and allows to pre-install packages using emscripten-forge and conda-forge.

Accessing the JupyterLite website#

After the build has completed, the site will be available on GitHub Pages. Go to to access it:

an animated screencast of using a JupyterLite website on GitHub Pages


By default the deployment provided by the jupyterlite/demo repo includes a .nojekyll file to bypass Jekyll processing on GitHub Pages.

See this blog post for more information.

Deploy a new version of JupyterLite#

To change the version of the prebuilt JupyterLite assets, update the jupyterlite-core package version in the requirements.txt file.

Commit and push the changes. The site will be deployed on the next push to the main branch.

Add additional requirements to the deployment#


The jupyterlite/demo repository uses a requirements.txt file to specify the dependencies. For demo purposes this file may contain extra kernels and extensions you might want to remove from your deployment. If that’s the case you can stick to a more minimal requirements.txt file such as:

# core package for building the JupyterLite website
# the Python kernel powered by Pyodide
# dependency for indexing the content


The requirements.txt file can be used to add extra prebuilt (also called federated) JupyterLab extensions to the deployed JupyterLite website. Follow the extension guide to learn more.

Using the xeus-python kernel and emscripten-forge#

Using a JupyterLite deployment with xeus-python, you can pre-install packages available both on conda-forge and emscripten-forge by specifying them to the environment.yml file.

By pre-installing packages, they are readily usable in the kernel and can be imported without the need for piplite.


You can add and update the default notebooks and files by clicking on the contents directory and dragging notebooks from your desktop onto the contents listing. Wait for the files to be uploaded and then save them (commit them) to the main branch of the repository.

Check out the how-to guide on managing content to learn more.

Further Information#

If you would like to customize your JupyterLite website, check out the different how-to guides.