Yeah, even from a non-accessibility standpoint, they're severely lacking. Text replacement shortcuts, mouse automation, it feels like a nightmare when you long for the control Windows users still have over their systems.
I've been using the Speed of Sound appimage for voice input, but it's just basic whisper to text on a global hotkey. I can get mouse control via python but only relative positioning. I tried to hack in some dead reckoning going to a bounding corner and coming back out but it's super cagey with accelleration which i'm not willing to turn off.
You're not an edge case; there are lots of us out there who would use a proper input api.