This project is a solution to a problem I was facing. I wanted to talk to the person at my door without using my camera’s two-way audio go2rtc, which was not working for me.
I tried several solutions with my Tapo 520WS camera, but I couldn’t get the two-way audio to work reliably. So, I built a simple alternative using Home Assistant and a Raspiaudio Muse Proto.
How it works
When someone rings the intercom:
-
I press a button in Home Assistant dashboard to record my voice as an MP3 file using the Voice Recorder Card: https://github.com/kukuxx/HA-Voice-Recorder-Card & sends the recorded file immediately to the Raspiaudio Muse Proto, which is connected to a 3W speaker near the door and the camera, so the person at the gate can hear me.
-
I can hear the person at the gate through the camera and communicate with them: they can hear my voice through the speaker, and I can hear their voice through the camera using https://github.com/dermotduffy/advanced-camera-card
Hardware
- Any type of camera installed at the gate.
- Raspiaudio Muse Proto, flash it with (ESPHome Media Players)
- 3W speaker.
- 9V → 5V step-down converter (LM2596).
- Wires & 3D-printed case (https://www.thingiverse.com/thing:5569337)
Wiring
I use the camera’s existing 9V power supply.
Pros
- No camera two-way audio required.
- Works with any type of camera.
- The camera lets me hear the person at the door.
- Hearing my audio again through the camera also confirms that the message was played.

- It can work locally over Wi-Fi and remotely if you are using a remote solution (Cloud, Duck DNS, Synology reverse proxy in my case, etc.).
Cons
- There is a small acceptable delay because the audio is recorded and saved before playing.
- It is not real-time communication, but it does the job.

Improvements and enhancements that can be made
- Intercom ring can be picked up by the Raspiaudio Muse Proto device.
- Use Raspiaudio Muse Proto built in MIC instead of Camera MIC for person on the gate.
- Reduce the small delay to make the communication more interactive.
- Use the existing intercom or camera speaker instead of a separate 3W speaker.
- Use a smaller DC step-down converter, such as the MP1584EN, instead of the LM2596.
- Make a smaller and more organized 3D-printed case.
Finally, it was an idea
, but it works well for my setup. Hopefully, it can help others facing the same problem.


