https://friture.org/ is the one I've been using.

Well, offering up my two cents there's something called Common Voice by Mozilla. I think it's like a voice dataset or something https://commonvoice.mozilla.org/en

You should use Signalizer, it's a really great FOSS Spectrum analyzer that I use all the time

I'm pretty sure there's a standalone version, but I always use it as a plugin, so if it doesnt have a standalone version I'd recommend loading it into Bespoke Synth, idk what bespoke does but it basically works flawlessly on any Linux audio system I've tried automatically, and it's just a fun software to mess with

Edit: a word

Vosk-server Vosk-api

Both are Apache-2.0 and the api is available on the AUR (albeit they're a bit outdated)

also here's a link to vosk's models

This might not be exactly what you're looking for but hopefully it provides another good option for you

This would be interesting, please send me a dm if you find anything. Good luck :)