If you can’t run a machine at home then you can’t self-host. You’re welcome to cloud-host though.

I’ve only ever added AI to n8n workflows through ollama nodes/tools and agents and whatnot.
I haven’t seen the version where an AI llm is baked in yet.
I did see we have tables now which is handy.
Like I said, I’m self self-hosted tho so, either way, don’t have to use it if you don’t want. I’d probably recommend anyone just install whichever is easiest unless, it’s forcing you to download the llm blobs/models and you don’t have the space for it.

2 versions? Nuts vs no nuts?
What do you mean? Your not being clear.
The AI starter kit is a docker stack, not a version or flavour of n8n.

Local hosted n8n: local AI llm agent (for privacy), that can use tools to search the web, check my emails and calendar, save memories, get Youtube transcript, etc.

Scheduled workflows to get me some stock info every morning.

Working on a research crawler with crawl4AI.

Only limitations seem to be my will to learn new stuff. I’m sure all this could be written in Python or something but I’m not a programmer.

Nah he’s right. Nobody buys a widget for 5 bucks and sells it for 10 bucks and says, “I made 10 bucks”. By your rational I could buy a car new for 25 grand, sell it 10 years later for 12 grand and say I “made” 12 grand off it.
“Make” has typically implied profit for as long as I can remember.

Same boat (in the learning cycle that is). No idea what immich is, but I got Stirling-PDF hosting in docker. I only learned the other day that localhost, is localhost for the container. I couldn’t get a bunch of stuff running for.ever, till I learned the way I was calling things needed to be to host.docker.internal.