Why is the Linux shell purely textual?

The other answers give cutting-edge workarounds.

Historically, the ancestor of Linux was developed in an environment where terminal standardization usually stripped out the marketable graphics features. Sound was single beeps, vocalized with ^G.

If it seems overdue to standardize a real-time character-encoded graphics protocol, you can look at the Tektronix 4104 and descendants.

That being said, typesetting was important early in UNIX. Even though the interface was command-line, UNIX implemented roff and maybe wouldn’t have succeeded if it couldn’t word-process into roff. Word processing was the first application for UNIX outside of the Labs.

Attempts have been made over the years at adding graphical capability: https://stackoverflow.com/questions/2114347/raster-graphics-...

You have a chicken egg problem; you need to get people to want to use both a different shell and a different terminal emulator, at the same time.

Textual mode fits well with all the tools. You can't grep an image sensibly.

Just like emojis can't transmit all the subtle nuances that words can, a textual interface is much more nuanced than a graphic shell.

In practical terms, a graphic shell can only give you what the shell's programmer has envisaged is possible, and has given you a menu to include those specific things.

That can't cover every single imaginable eventuality. If you don't believe that, try making a menu on paper that covers everything possible that a textual shell can do.

There's nothing stoping someone from making a Linux terminal that can contain images. The problem is you'd need to make a shell that's aware of how to wrap text around images, and you'd need a terminal app that can display them. And then you need terminal apps that know how to send images into this system.