10 or so years ago when Minecraft was coming out of beta, I was a precocious enough child to run a server on some spare hardware lying around the house. (Realms wasn't a thing until like 2014)
Good times were had between my friends and I, but that all changed after I started becoming intoxicated by the power. Whoops.
Give it a try, but I advise against giving a literal child the ability to ban people. I would know.
The kids usually start a voice FaceTime group chat and then start up Minecraft, which keeps the voice chat going in the background. It works great. (If they are playing with someone that doesn’t use FaceTime: I’m pretty sure that the Facebook Messenger for Kids app also works)
I’ll also put in a plug for the Apple Watch cellular - we have them for the kids using the family setup. The kids then have a phone without having a phone, they have the blue iMessage bubble in chat, etc. But no camera, no bad social apps, etc. And you can have them delegate management of their address book to your phone and then set the watch to ignore everything except contacts in the address book. A side benefit is that each watch is $10 a month instead of the $30-50 per month a real phone costs.
Edit: reading the rest of your comment I see that this doesn't exactly meet your 'voice and video' specifications. Apologies.