Screen calls (voice to text)

Since iOS and Android now both offer this feature, it would be nice to have it with Linux phones like the Librem 5 as well. What I mean by screen calls:

You get called by another person and you might not be sure who’s calling you or you don’t want to talk to them. So the software can answer the call via a button and show via voice to text trans-coding on the screen what the other person is saying to let you decide whether to talk to them or hang up.

This should definitely be possible in theory. There’s a library called april-asr availabe for Linux that can convert speech to text in real time. It works quite decent with English already. It would also be possible to have LiveCaptions open in parallel and switch off your microphone but I personally think an integration in the Calls app would be much more convenient.

Also since we are on Linux and modular approaches are awesome. Maybe once the library is integrated, there might be the possibility to let it interact with custom scripts to let a bot handle some phone calls for you. This would just need text-to-voice synthesis as well. But it would make waiting in a phone call queue so incredible convenient to have a bot answer basic questions for you or navigating a request to let you know when exactly a person is on the other end…

4 Likes