I am new to this stuff, so please have patience with me.
My first win became to successfully make a local install of n8n in a Docker environment on my old MacBook Pro M1 and make it access a locally installed LLM instead of reaching out to any paid cloud edition.
Then I did a similar thing on my PC but there I chose to install n8n without Docker to try a different environment. After that I made n8n connect to a LLM on a local LM Studio server environment, just to see if it works.
By running different setup variations I can get more experience in what can go wrong and discover which limitations these setups have.