I'm not sure if I will get around to writing it up, but last night, I built on this article, tried out LangChain, and wrote my first few tools to expose.
In my first demo, I had Ollama + Llama3.1 calling multiple python functions which was amazing.
By the end of the night, it would only make one custom tool call and I have zero clue how to fix it or know if it's even fixable.