LocalAI integration with Nextcloud AIO (30.0.8, Hub 9) – looking for updated guide and advice
Hi everyone,
I’m currently running Nextcloud AIO (version 30.0.8 with Hub 9) on an Ubuntu Server 24.04 machine with a 6-core CPU + 6 threads).
About half a year ago, I tested LocalAI on this same server as a Docker container. Unfortunately, I didn’t manage to get the workers running properly back then, so the performance was poor, and it wasn’t really usable.
As a temporary workaround, I connected Nextcloud to the OpenAI API, and today I successfully configured 4 workers, which are now working very well – I’m getting AI responses within 2–3 seconds.
To be honest, I’m not entirely sure whether the worker setup I’m using now is exactly the same as what LocalAI would require, but since it works perfectly with OpenAI, I assume it’s not a major issue.
That said, I’ve noticed several new LocalAI-related apps in the Nextcloud App Store recently, and I’m really interested in switching back to a self-hosted AI setup for more control and privacy.
So I’d like to ask for advice, updated instructions, or any reliable guide on how to properly run LocalAI with the current version of Nextcloud AIO (30.0.8, Hub 9), ideally in the simplest and cleanest way possible, using Docker.
Any help, experience, or link to an up-to-date guide would be greatly appreciated!