Post

This is especially annoying if you want Discord or Slack to always output on your headphones, and a music streaming page on a different output. The Linux audio system supports this, but since everything just ends up as "Chromium", all new audio is always routed to the last selection.

I'm not a C++ developer, but I wanted to fix that. And is there a better usage of morally questionable AI models than to actually improve open source code?

So now I have a local fork of ungoogled-chromium that actually sets the correct pulseaudio properties for each audio stream. It dynamically updates the stream label when the document title changes and it ensures that the correct app information for web apps is set.