Plugins + Scripts VoiceControl2

[StyleMavin]

StyleMavin submitted a new resource:

VoiceControl2 - Standalone voice control based on VeeRifter.VoiceControl using Vosk, not Windows Speech Recognition

View attachment 601245View attachment 601248
VoiceControl2 — hands-free voice triggers for VAM

Trigger anything in VAM with your voice. Speak a keyword or short phrase and fire any VAM trigger — load looks, start animations, switch scenes, toggle plugins, whatever you can wire to a trigger.

This is a modernized successor to the classic VoiceControl / VoiceTrigger plugins, rebuilt to survive Windows updates...

Read more about this resource...

 

[Dlesser]

Hi Buddy! Great plugin. I’ve already used it as a base to build voice communication and connected LLM control through LM Studio to my sex machine.

I’m also working on the backend side in Python now, so the model can respond with a real, nice-sounding voice. Once I test the Python part properly, I’ll move the C# part into one unified program.

 

[StyleMavin]

Hi Buddy! Great plugin. I’ve already used it as a base to build voice communication and connected LLM control through LM Studio to my sex machine.

I’m also working on the backend side in Python now, so the model can respond with a real, nice-sounding voice. Once I test the Python part properly, I’ll move the C# part into one unified program.

Thank you! I’m glad you like my plug-in. Let me know how your project progresses. I would like to see the results.

 

[Dlesser]

Thank you! I’m glad you like my plug-in. Let me know how your project progresses. I would like to see the results.

Hi again. To be honest, I have already done almost everything. I reworked your code, and now the script connects to LM Studio, so you can talk to the model by voice. The model also answers with a human voice.

However, I had to make the second part in Python and connect everything into one UI. I used XTTS in Python, with a model where you can give it any audio file as an example — even just sounds or moans — and the model answers in the same voice, with intonation that is almost impossible to distinguish from a real human voice. It sounds especially cool if you give it just moans as the reference; the replies sound almost like phone sex))) The voice is so alive that it is almost impossible to tell it apart from a real one.
A little later I will publish an update both for the sex machine and, I think, a separate communication script as well, so it can be used separately from my main scripts. But that will probably be around the end of the weekend.
What is convenient is that I simply give voice commands to the LLM, and it does what I need in the game. As you understand, any language can be connected, and the replies can also be in any language.
Right now I have also connected 3 cameras. I installed one on the character’s head, around the eye area, so the LLM can see everything in the scene, and the others are placed from the side so the LLM can see the scene itself.
Now I am finishing the program so it can work both through the CPU and on the GPU. I also still need to clean up the server UI, fix small things, and make sure no service garbage gets into the replies.
All these small things will probably take me until the end of the week, while I connect everything to my sex machine, so the LLM can control it and see the process in the game.

To be honest, I haven’t slept for a couple of nights, so I’m really exhausted. I need to take a couple of days off, since the main work is already done.
But I think this will be better than https://hub.virtamate.com/resources/koboldlink.49327/, because in my case the model replies with a real, living human voice.

 

[StyleMavin]

Hi again. To be honest, I have already done almost everything. I reworked your code, and now the script connects to LM Studio, so you can talk to the model by voice. The model also answers with a human voice.

However, I had to make the second part in Python and connect everything into one UI. I used XTTS in Python, with a model where you can give it any audio file as an example — even just sounds or moans — and the model answers in the same voice, with intonation that is almost impossible to distinguish from a real human voice. It sounds especially cool if you give it just moans as the reference; the replies sound almost like phone sex))) The voice is so alive that it is almost impossible to tell it apart from a real one.
A little later I will publish an update both for the sex machine and, I think, a separate communication script as well, so it can be used separately from my main scripts. But that will probably be around the end of the weekend.
What is convenient is that I simply give voice commands to the LLM, and it does what I need in the game. As you understand, any language can be connected, and the replies can also be in any language.
Right now I have also connected 3 cameras. I installed one on the character’s head, around the eye area, so the LLM can see everything in the scene, and the others are placed from the side so the LLM can see the scene itself.
Now I am finishing the program so it can work both through the CPU and on the GPU. I also still need to clean up the server UI, fix small things, and make sure no service garbage gets into the replies.
All these small things will probably take me until the end of the week, while I connect everything to my sex machine, so the LLM can control it and see the process in the game.

To be honest, I haven’t slept for a couple of nights, so I’m really exhausted. I need to take a couple of days off, since the main work is already done.
But I think this will be better than https://hub.virtamate.com/resources/koboldlink.49327/, because in my case the model replies with a real, living human voice.

That sounds pretty dope! How powerful is the GPU needed to run this model and get good performance? The part about giving it moans as a reference made me imagine some hilarious voice replies and I started laughing. I tried something like that with ElevenLabs a while ago and it was funny by not useful.
Post some examples of these interactions if you can. It sounds really interesting, especially the part about telling the LLM what you want in game and the LLM making it happen.

 

[Dlesser]

That sounds pretty dope! How powerful is the GPU needed to run this model and get good performance? The part about giving it moans as a reference made me imagine some hilarious voice replies and I started laughing. I tried something like that with ElevenLabs a while ago and it was funny by not useful.
Post some examples of these interactions if you can. It sounds really interesting, especially the part about telling the LLM what you want in game and the LLM making it happen.

It is not translated into English there yet. I work in my native language, and usually translate everything at the end, when I publish it. So I do not want to send the video here yet.
If you’d like, I can open the GitHub project for you. You will be able to download it and, if you want, remake it however you like.
I am not chasing ownership rights; I just like making interactivity. You can find me, for example, on Discord, where we can talk more freely.