Just came across this video <https://www.youtube.com/watch?v=egrMopDIvPE> on the “Learn Linux TV” channel, talking about this new device that does text to speech. The actual speech-synthesis software is open-source. The original “Mimic1” has been around for a while, but the new version takes advantage of newer Raspberry π hardware. He talks about it not needing an Internet connection any more: I think this was a requirement of the old hardware, not the software as such. You can build and install the software (old or new) yourself, it’s freely available from their GitHub repos. Interesting that the build instructions for Microsoft Windows recommend doing a cross-build from Linux, this being faster than a native Windows build.