I propose to implement this using either Python + Flask (micro HTTP framework) or C / C++ (e.g. using the Pistache C++ REST framework). Would you object to either of those solutions? I propose using a recent version of Raspbian Lite for the SD card image.
I have a Raspberry Pi model 3 to test and develop with. I have used SystemD for automatic program startup on boot previously. There are probably other options also.
I have worked with audio on Raspberry Pi on projects in the past - with ALSA, OSS and OpenAL. (In this case, handling audio may be as simple as using `aplay`).
PLEASE NOTE: I may need to adjust my bid slightly based on further details you provide. In the meantime, please regard this bid as (at least somewhat) PROVISIONAL.
I am based in Ireland.