- You write a .import config file which is essentially a Python list of importer objects configured with your account names
- You run bean-extract with it and with some directory when your download stuff from your bank/credit cards
- bean-extract looks at all the files in that directory and identifies which file it to be processed by which importer ("identify")
- bean-extract then runs each of the matching importers on them
- you redirect that output to a file
- you then manually eyeball (they like to say "reconcile") for correctness, complete ("categorize") and copy paste the imported transactions in the relevant part of your .beancount file, in the manner that suits you best