I think @dcreater was asking how these various structee decoding providers compare with how pydantic ai handles structured output, i.e via tool calling, forcing the LLM to use a tool and its arguments are a json schema hence you read the tool call arguments and get a structured output.
What you mentioned is pretty much the way to go. You work on your config on the weekend what you thought was missing throughout your week. I wish everyone tells that the people starting out with vim/Emacs.
Okay this is irrelevant to the discussion but I just typed in `github/mevdschee/2048.c` in the address bar expecting Firefox to perform a Google search but it instead lead me to the repo! What!