You can do CW entirely in software without learning it. I've been considering trying this out because the lower bands that technicians are permitted to use here in the US are CW and not voice.
Most of the software that decodes CW isn’t that great unless you have a consistent S9 signal and the keying is even (IE, not straight keyed). Once the signal wanders or is even partially in the noise the software will just spit out gibberish but a human would still easily be able to copy it.