Voxta AI

[RussellRick]

Is the lesbian scene with the red demon/succubus available here on Hub? I tried finding it, but had no success.

 

[Jacob42]

Purchased a month of this to try it out. My goal was to do everything locally since my system can handle it. However I ran into major issues with the text to speech option (Whisper). For one, it doesn't pick up when I'm speaking half the time. So it'll cut off my sentence mid sentence even though I am clearly still speaking (even when I add a 2 SECOND delay to the time out). For two, even when there's completely silence, it'll pick up completely random works like "hello" for no apparent reason. I have no issues with regular dictation software (like build in windows one), so it's not my mic. I also tried several different Whisper models (large, small, turbo, regular, etc) but they're all basically the same as far as I can tell from the few I tried.

Also other people don't seem to be reporting the same issues with Whisper as me and Whisper is touted by many as the current best open source option for TTS. So, I really feel like it has to be an issue with Voxta itself.

Tried messaging creator on Patreon and got one response saying to try different models (which I had already explained I did) or try using an alternative TTS option (none if which are local!) and that it's something due to with my setup. But that doesn't make any sense because everything else with Voxta seems to work fine, and like I said above other TTS through Microsoft or Google with my mic are picking up everything 100%. I tried messaging him again suggesting I think it's an issue with Voxta itself and haven't gotten a response in a month. Just had my sub renew too because I forgot to cancel it, ugh.

Don't think Voxta is well polished for local use. And I don't think VamX plugin support local models either. So as of now there's really no good options for local models I can find that work with Virt A Mate. If you're looking for local use, I would skip this mod.

 

[BillTheButcher]

Purchased a month of this to try it out. My goal was to do everything locally since my system can handle it. However I ran into major issues with the text to speech option (Whisper). For one, it doesn't pick up when I'm speaking half the time. So it'll cut off my sentence mid sentence even though I am clearly still speaking (even when I add a 2 SECOND delay to the time out). For two, even when there's completely silence, it'll pick up completely random works like "hello" for no apparent reason. I have no issues with regular dictation software (like build in windows one), so it's not my mic. I also tried several different Whisper models (large, small, turbo, regular, etc) but they're all basically the same as far as I can tell from the few I tried.

Also other people don't seem to be reporting the same issues with Whisper as me and Whisper is touted by many as the current best open source option for TTS. So, I really feel like it has to be an issue with Voxta itself.

Tried messaging creator on Patreon and got one response saying to try different models (which I had already explained I did) or try using an alternative TTS option (none if which are local!) and that it's something due to with my setup. But that doesn't make any sense because everything else with Voxta seems to work fine, and like I said above other TTS through Microsoft or Google with my mic are picking up everything 100%. I tried messaging him again suggesting I think it's an issue with Voxta itself and haven't gotten a response in a month. Just had my sub renew too because I forgot to cancel it, ugh.

Don't think Voxta is well polished for local use. And I don't think VamX plugin support local models either. So as of now there's really no good options for local models I can find that work with Virt A Mate. If you're looking for local use, I would skip this mod.

It is a decent plug-in idea and it definitely needs polishing, I also run local, I have never used TTS because I always disable it and type anyway, yeah it sucks, I have seen some changes through out the updates that I do not like either, some updates work great and others not so much! It's like voxta is injecting guidelines or filters in, I have models that are completely uncensored made for porn or adult erotica and have tried all them with voxta, but voxta always injects something about boundaries and respect now. They used to work great, but I'm not gonna be lectured about boundaries and respect if I use a test question like can I smell your feet for example!...NSFW means NSFW. Yet I could literally put the model in a dungeon with a ball gag in her mouth wtf!....or put her in chains, so it's almost like they are injecting some kind of political agenda into the app now, I have spent many hours with chatgpt on trying to figure this out and that is what it has concluded. Good things always come to a end I suppose.