Velesio AI Server is a web application aiming to provide a simple framework to host open source AI models, be it on-prem or in the cloud, which can be used by both Unity and Web Applications.
To access this in unity you can use the linked repos bellow, a unified unity integration for all of this app's services is currently under development and will soon be published to the Asset Store.
This has only been tested on NVIDIA GPUs with Cuda 12.2 and higher and is developed with that environment in mind.
Deploy effortlessly using the Runpod template. It's a plug-and-play solution for hosting in the cloud.
New to Runpod? Use my referral link to get started!
Get up and running locally with just a few steps:
-
Install Docker Compose
Follow the official guide to set up Docker Compose. -
Clone & Compose
git clone https://github.com/Velesio/Velesio-AIServer.git
cd Velesio-AIServer/velesio
docker compose up -dExposed Ports:
3000- Vite frontend1337- Undream AI server7860- Stable Diffusion Automatic 1111 webserver
These ports are reverse-proxied through Nginx. Access control can be managed via the ALLOWLIST environment variables, as Runpod currently lacks built-in firewall functionality.
- ⚡ Plug and play AI hosting for Unity — no heavy setup required
- 🧠 Built for self-hosting and the cloud
- 🛠️ Built to be open, extensible, and developer-friendly
Stay tuned for:
- ✅ Unity Integration Asset
- ✅ Documentation
- ✅ API Service
- ✅ Ollama support
- ✅ Flux support
Follow the repo, drop a ⭐, and join the journey!
Veles would not be possible without the amazing work from the open-source community. Special thanks to:
-
🧠 LLM for Unity by Undream AI (MIT License)
A streamlined interface for integrating LLMs into Unity. -
🎨 Stable-Diffusion-Unity-Integration by Dobrado76 (Apache 2.0)
A powerful image-generation pipeline with Stable Diffusion.
Their projects serve as a foundation and inspiration for this unified framework.