[chasd00]

If you want a code centric interface it's pretty easy to use llangchain with local models. For example, you specify a model from hugging face and it will download and run it locally.

https://python.langchain.com/docs/get_started/introduction

I like llangchain but it can get complex for use cases beyond a simple "give the llm a string, get a string back". I've found myself spending more time in llangchain docs than working on my actual idea/problem. However, it's still a very good framework and they've done an amazing job IMO.

edit: "Are options ‘idiot proof’ yet?" - from my limited experience, Ollama is about as straightforward as it gets.

[ShamelessC]

Why are you using two l’s to spell langchain?