Home Creators Posts Import Register Favorites Logout

Videos

  • Grabación de pantalla 2026-11-18 044055.mp4
  • Grabación de pantalla 2026-11-18 064644.mp4
  • Grabación de pantalla 2026-11-18 065058.mp4
  • Grabación de pantalla 2026-11-18 070243.mp4
  • Grabación de pantalla 2026-11-18 070732.mp4

Downloads

Content

I've been working on a chat for a local AI,well, actually, Gemini is writing the code 😅. I'm doing this to create alternatives to local servers and to use LM Studio on my phone or tablet. I guess it should work on other devices in your local network with a browser too (sorry to those of you with a PS5 🤣).


And it occurred to me that you could use it in Vam,in a way, unfortunately, the default web browser in Vam doesn't work, so I have to use an external one. Lip-syncing is handled by the SystemAudioAction plugin from Pluginidea, and although it's not necessary, you can use the E-motion plugin from VrAdultFun to make it feel more lifelike with micro-movements and facial expressions. You can also create your own idle animations.


The app is designed to work just like a normal chat, but it has certain features, like the Lorebook, so that the models can remember important details from the conversation, and text-to-speech with the option to use either Google’s version or Edge’s ,the latter sounds more natural.

I asked him to make it like a YouTube interface or a web player, so I could set custom backgrounds or animated avatars, you can turn this feature on and off.

It depends on your local model; it can analyze images and respond. In my example, I'm using Gemma 4 with 4B parameters.

I know there are better options, but as you well know, I like to make my own things for my personal use and not rely on others and their tastes

It's still in the planning phase, and there are features yet to be added, such as using an API key to access cloud-based AI and speech-to-text (I won't be using this feature, but some people might find it interesting). This app isn't just for VAM; it's also for RPGs or anything else where you want conversations or guides.

My PC is way too old,a 5600X, 32 DDR4, and an RTX 2080S, but how the AI runs depends entirely on the model you choose in LM Studio. On a Gemma 4 in 4B, I get 100–120 tokens/s without any load from Vam or games; with load only in RAM, I get 10–15 tokens/s.

Video demos are available to members.

English voice

Spanish voice

Spanish voice speaking englsh

You can regenerate audio

Male version

I don't know why, but the Colombian Spanish version generates voices really fast. Maybe if you want to role-play with a Latina with an accent, like certain wife in a modern family.

Files

Comments

No comments found for this post.