Local AI Integration

[BSweet]

I could imagine that it would be rather straight forward to load a Docker container onto a section of the NAS. I would LOVE to see some kind of optimization of an AI assistant if I installed an app on HexOS. So far thinking I would love to pair HexOS with Homey to have a completely offline Smart Home solution. Then throw in devices such as Reolink which an integrate with local NAS networks (Reolink has a Synology app even), and I would be able to host a local AI to know about all my documents, code, give it access to camera feed, speakers, lights. Soo many possibilities. But if HexOS could regognize a GPU as dedicated to an AI for all of this, that would be a massive reason for AI to integrate so powerfully on this platform.

Local private AI with intimate information, and still massive speed. Just start with a downloaded model.

 

[kobypierce]

This sounds great, I have sometime like this running on a Proxmox VM right now on an AI VM server I built. It has 4x 3060 12GB in them. It's running Ollama, and Open WebUI. I saw in the Linus Tech Tips Video something about Ollama I thought, but it says there are only two apps so far so we'll see when i get access if i can transfer it over to HexOS.

[ubergeek]

I did see that open web ui and ollama were in the App list., or were you looking for something else?