# CodeBaby and Using Babylon.js

**URL:** <https://forum.babylonjs.com/t/codebaby-and-using-babylon-js/60711>\
**Category:** Demos and projects\
**Created:** [September 29, 2025, 5:43pm UTC](https://forum.babylonjs.com/t/codebaby-and-using-babylon-js/60711 "2025-09-29T17:43:55Z")\
**Posts on this page:** 1\
**Showing post:** 2

<div class="post-metadata">

**Author:** ![Tricotou](https://sea2.discourse-cdn.com/flex024/user_avatar/forum.babylonjs.com/tricotou/32/5844_2.png) [@Tricotou](https://forum.babylonjs.com/u/Tricotou)\
**Post date:** [September 30, 2025, 1:45pm UTC](https://forum.babylonjs.com/t/codebaby-and-using-babylon-js/60711/2 "2025-09-30T13:45:52Z")

</div>

Hey @alexa welcome !

> [@alexa](#):
>
> We’re a live-rendering animation company that combines AI with real-time avatars to enable interactive conversations. Here’s one of our concierge demos in action:[https://concierge.codebaby.com/concierge/hotel/](https://concierge.codebaby.com/concierge/hotel/)

I had a quick test, it looks nice 🙂  
A few questions about this

- What kind of LLM do you run behind those avatars ? Are they 100% custom, or are you based on a tier-AI such as ChatGPT, Grok, Llama3.2, Gema3, etc… ?
- Are the 3D animated gestures deep based on the LLM output, or hard coded and limited in diversity ?

* * *

I also did some tests animating Webbased avatar using BabylonJS, have a look 🙂

> [@BabylonJS + Deep-Learning](https://forum.babylonjs.com/t/babylonjs-deep-learning/57634/6):
>
> I’m so impressed by how well BabylonJS handles multiple morph targets in realtime, flawlessly ! smiling_face_with_three_hearts [multi\_pose] [multi\_mouth] [multi\_eyes]

++  
Tricotou

---

_[View the full topic](https://forum.babylonjs.com/t/codebaby-and-using-babylon-js/60711)._