Browser Functions: Serverless Where the Runtime Is a Browser

Serverless functions are a familiar model: deploy a snippet of back-end code such as Node to the cloud and call it over a URL. A concept described by Richard Young flips the runtime — instead of a back-end language answering that URL, an actual browser does.

The project, browser functions, is framed by its author as a new serverless platform built on web browser execution engines. The pitch is that full stack development can now be done with Web APIs alone:

  • Networking and files: the fetch API for reading and writing network resources.
  • Caching: localStorage for storing data.
  • Image processing: a CSS filter on an img tag to blur an image.
  • Sessions: cookies for session management.
  • Concurrency: Web Workers for multi-threading.
  • Raw speed or other languages: WebAssembly for natively compiled performance or a language other than JavaScript.

Because browsers have grown so capable, leaning on the tasks they already handle well is a sensible move — and an inventive one.

Examples and source code live in the project's GitHub repository; the write-up is at the direct link above.