I'm looking to see if someone could point me in the right direction in terms of understanding what's needed for to host a backend server for the new Push To Talk API. Additionally, I'm curious to know what type of latency times people are getting. We've ran a mumble server which unfortunately had a considerable amount of latency.
The PushToTalk framework allows your app to use any network protocol and audio encoding process you’d like. The only backend requirement is that your server must use Apple Push Notifications and the new pushtotalk notification type to alert your iOS app when a new incoming message is ready to be played. Review the “Receiving Audio” section on the Creating a Push to Talk app page for more information.
Support for Push to Talk messages was added to the Local Push Connectivity framework in iOS 16.4 if your backend service is not able to Apple Push Notifications: