Even on my 8 cores laptop it is slow to generate response compared to my computer with GPU. NAS CPU is just not powerful enough to run AI.

Ollama can be installed easily on Windows with it's setup file. Go to ollama.com and download the setup file for Windows. After install open command prompt or power shell (might need to run as admin) and type "ollama run <insert model name>"
It will then download the model. Better to get llama3:instruct as this have been tweaked to give better answer. Now you can type anything to ask the AI. phi3: instruct also another good model to try.

If you want to have UI (user interface), install this
https://github.com/open-webui/open-webui

I recommend using the docker method. You still need to install ollama, this is just create web interface and also for RAG. After install docker, just run this on command prompt


You can attach text doc or insert link and ask questions about it. With model such as llava you can also attach picture and ask about it, but this is still very infancy.

Once done, you access it on http://localhost:3000
With some configuration on your computer firewall, you can also access the web ui with any device connected to the same local network, your computer become the AI server for your household.