I think there should also be a custom separator for each field and between each card as well… As I am only able to get the first card converted, but the rest are not properly separated by pipe or semicolon
However, I see that there might have been some error issues, which makes it more understandable.
Quoting Dae from the Github thread:
The extension requirement is so that Anki can identify what type of file is being imported. The previous approach lead to confusing errors when users tried to import unsupported file types, as Anki was trying to import them as a CSV.