i want recommendations for free text to speech interfaces like those used for speech impaired people (screens of icons for quick vocabulary, customizable, with option for full text input)
bonus points for something with a highly customizable voice, or also seeking recommendations for any tts voice that's local, customizable, not genAI, runs on old hardware. i don't want solutions that only work with the system's black box TTS i have to find arcane packs for
please boost! 
@mavica_again
Have you tried espeak? That one still sounds fairly synthetic.
@Kavus i remember trying it some ~10 years ago and it not producing any output (due to my not figuring out how to set it up properly as i seem to remember the documentation at the time being very academic and requiring prior knowledge)
i'll take a look at it again but i'm hoping for raw voice synthesizers rather than something that uses pre-recorded voice databases as a source
@mavica_again
Huh! I didn't know it needed any special setup, I just install it and it goes.
@Kavus i just remember it needing voice files it didn't provide, nor them being easy to being found. MBROLA has gotten a lot more commonplace since the last time i tried it
@mavica_again
Ohhhhh… I think they just come with it now? I can check when I get home.
@mavica_again so, I've got a potential solution, but I'm getting caught up on some developments and wanting to explain some details.
Short version: Piper TTS https://github.com/OHF-Voice/piper1-gpl
@mavica_again It was originally called Mimic3, made by a company called Mycroft.ai even though they weren't an AI company in the way that genAI is right now. They focused on language processing, in order to make a local voice assistant. Required figuring out what people were saying, what they were meaning, and a way to reply.
That lead to Mimic3, their library that used university voice samples to make models based on speakers, and while they sound high quality, they still sound artificial.
@mavica_again The company ended up getting sued into bankruptcy by a patent troll, but they were able to have their work picked up by what looks like maybe the folks who do Home Assistant? And, they relaunched the Mimic3 project as Piper TTS, and have been working on it from there.
Hopefully, it'd be what you're looking for, and might run well enough on lower end hardware to fit your needs.
@gothpanda that's interesting and i'll take a look at it but i'm looking more for voice synthesizers than anything that uses recorded speech as a source
that could've been clearer in my first post
thank you
@mavica_again there is this company making emulations of old tts engines for musical applications. They have what looks like a free trial only though.
@jaseg thanks, aware of it but the price is prohibitive for me atm, and i'm not sure if the VST format will be easy for me to integrate into anything but a musical DAW
yes the idea is to find voices that sound robotic, synthesized
i know that's not what most people using TTS want them to sound like, that's why it's so hard for me to find what i want
yes i already know about dectalk. i know there are many other TTS engines from the 80s and 90s and those all feel like vapourware now because you can't find anything on them anymore