CodePen vs JupyterLite

CodePen is a web-based social development environment for creating, testing, deploying, sharing and discovering front-end code. In a Pen, users write HTML, CSS and JavaScript and see a live browser preview. It supports HTML preprocessors such as Markdown, Haml, Slim and Pug, and CSS preprocessors including Sass, Less, Stylus and PostCSS. The CodePen 2.0 editor supports automatic preprocessing, live previews, file systems, version history, collaboration and one-click Pen deployment. Users can add external CSS and JavaScript resources, reuse other Pens, find libraries through CDNjs and import bare npm packages through its compiler. Pens can be embedded on other websites, and CodePen documents a REST API and other embed and prefill options. Pens and Projects are public by default. PRO users can make Pens and Collections private, but anyone with the shared URL can view them without logging in. The free plan includes unlimited public Pens and three files per Pen. Starter costs $8 a month, Developer $12 a month and Super $26 a month. CodePen does not offer a self-hosted version.

JupyterLite is a free, open-source JupyterLab distribution that runs entirely in a browser. It provides JupyterLab and Jupyter Notebook interfaces, plus a minimal browser REPL. Documented Python kernels use Pyodide or Xeus Python, and contributors maintain an R kernel based on Xeus R. Interactive visualization support includes Altair, bqplot, ipywidgets, matplotlib, and Plotly. Users can create, edit, save, upload, and download files in the browser, where files are stored by default in IndexedDB. Data stays there unless an enabled extension sends it to an external service or notebook code uses the browser’s fetch mechanism. JupyterLite can be served as static HTTP(S) assets, embedded in another application, or deployed to supported static hosts without a dedicated application server. Browser storage is limited, though the documentation says smaller uploads up to about 50 MB should be fine. Some Python packages may not work because WebAssembly kernels cannot use certain native extensions, system libraries, threading, or filesystem access patterns. Where supported, a JupyterLite site can be installed as an application. The project says only its two latest releases are actively supported.