Johan Mabille, Thorsten Beier Xeus kernels in the browser | JupyterCon 2023

"Learn how Xeus kernels in the browser revolutionize interactive web applications with Jupyter Lite, a new way to run kernels without a separate server, and explore its advantages over traditional kernels."

Key takeaways
  • Jupyter Lite is a new way to run Jupyter kernels in the browser, without the need for a separate server.
  • Ksus is a custom implementation of the Jupyter kernel protocol, written in C++ and compiled to WebAssembly.
  • Zeus Lite is a subset of Ksus, optimized for smaller kernels and faster startup times.
  • Jupyter Lite can be used to create interactive web applications, without the need for a separate server.
  • The browser acts as the server, and the kernel communicates with the browser using WebAssembly.
  • The kernel can be written in any language that has C or C++ bindings, and can be compiled to WebAssembly using Emscripten.
  • The GZOOS cookie cutter is a template for creating new kernels, which includes the necessary code and infrastructure.
  • Jupyter Lite is portable and can run on any browser that supports WebAssembly.
  • The kernel can be extended with custom code and modules, and can be used to create complex interactive applications.
  • Jupyter Lite has several advantages over traditional Jupyter kernels, including faster startup times and smaller kernel size.
  • The kernel can be used to create interactive web applications, without the need for a separate server.
  • Jupyter Lite can be used to create interactive web applications, without the need for a separate server.
  • The browser acts as the server, and the kernel communicates with the browser using WebAssembly.
  • The kernel can be written in any language that has C or C++ bindings, and can be compiled to WebAssembly using Emscripten.
  • The GZOOS cookie cutter is a template for creating new kernels, which includes the necessary code and infrastructure.
  • Jupyter Lite is portable and can run on any browser that supports WebAssembly.
  • The kernel can be extended with custom code and modules, and can be used to create complex interactive applications.
  • Jupyter Lite has several advantages over traditional Jupyter kernels, including faster startup times and smaller kernel size. *