Using Mycroft with a voice scrambler


I’m thinking about the possibility to add some kind of a voice scrambler to my mycroft project. Has anyone tried anything like this? Basically I need to use Google could services for STT, since it has the best language support for the language I’m building for, but I would like the recorded voice to be unidentifyable (GDPR - yay!)

Short of buying a Darth Vader mask from the local toy store and speaking throu that, does any one have any ideas of where to start?

Build a deepspeech model for yourself instead?