Cutsom doorbell, 2 way audio

Hello community !

I have the goal to speak with anybody ringing the doorbell.

I used a Ring doorbell, but the integration in HA was not very well, and the app is limited if you don’t get the subscription plan.

So I’m going the hard way. In fact at the front of my house, about ~30m I have the electric gate/portal, the mail box the camera, and the doorbell.

Currently i’m using a Raspberry pi GPIO’s with remote_pi, PoE powered. Why ? Because it don’t use the wifi and it give me a lot of capabilities of Input and Output :

Output :
-Open the gate for car
-Open the gate for pedestrian
-Switch on the light at the gate

Input :
-Mailbox pane
-Mailbox door (so I know if i get a letter or a package)
-Doorbell button
-Current loop (for going out with a car)
-Gate sensor (open/close)

And with that a Reolink RLC-520, PoE too, for motion and plate detection.
The RLC-520 has one-way audio. I was thinking to switch for a RLC-820A that have 2 way audio, but it seems complicated/impossible to use a 2 way audio of a camera in HA like I used to with the Ring Doorbell.

Listeting the audio of camera is not really a problem.
The final goal is something with 2 way audio with a nice integration.

But the first step would be a TTS for giving indication to someone at the doorbell. As I already have a Pi at this location, is it possible to do that, running raspbian or something ?
Or otherwise, configure a HA instance on that pi with a speaker plugged to it, and use Remote HA ?

Thanks !

Look at SIP server for two way audio. Asterisk or freePBX

Also consider Dahua VTO and (I think DTO) doorbell camera and screen. No cloud needed, two way doorbell out of box, ability to two way audio in HA(need SIP Lovelace card), also sip server built in so need to run that (which to me is a pain).

RasPi is perfect otherwise