How to set up Local AI

Run LLMs on your own hardware. This guide covers how to access it, log in for the first time, configure the basics, and where its data lives so you can back it up.

1. What is Local AI?

Run large language models locally with Ollama, and chat with them through Open WebUI — a private, ChatGPT-like interface. No data ever leaves your server.

2. Before you start

LLMs need significant RAM. 3B models run on 8GB, 7B on 16GB, 13B+ needs 32GB+. The UGREEN DXP2800 with 16GB RAM can comfortably run 7B models.

3. How to access it

Once it's enabled, Local AI runs on your SparkBox at:

http://<your-NAS-IP>:8090

Create your first account after opening the web UI — that account becomes admin

4. Settings you can change

You can set these when you enable the module, or later from Apps → Local AI in the dashboard:

• Default Model — Which model to download by default. Small models (3B-7B) work on limited hardware. Large models (70B) need 64GB+ RAM.
• Open WebUI Secret Key — Configurable value.

5. Where your data lives

Local AI's config and persistent data are stored on the SparkBox host under:

/opt/sparkbox/modules/localai/config/

Backups via the SparkBox dashboard's Backup tile include this directory by default, so a single restore brings Local AI back to its current state on a fresh install.

Next steps