I successfully switched to ledger a few years ago after a few failed attempts years earlier. I attribute success to: https://github.com/quentinsf/icsv2ledger ht

I successfully switched to ledger a few years ago after a few failed attempts years earlier. I attribute success to: https://github.com/quentinsf/icsv2ledger https://github.com/quentinsf/icsv2ledger It is the transaction import script that has really worked well for me without quirks. A lot of people recommend "reckon"--which I've tried--but I found that while Bayesian scoring sounds good, a solid set of regex work better. icsv2ledger does a great job of allowing you to build up a pool of search terms to map inputs to categories. And since the regex set is an artifact that you can tweak and add to, you end up with a very stable and comprehensive import process. I also like that it will (optionally) add the raw CSV that was imported as a comment in the transaction, and use that to avoid reimporting transactions. While this is a little visually noisy in the ledger file, I'm blind to it by now and it makes it rock solid to reimport CSV files without fear of duplicates. This also provides a neat way of dealing with each side of a transaction coming in separately. For example, both my checking and credit card list the transaction that corresponds to me paying my bill, but that is only one transaction in the ledger file. After I've done the import of everything, I simply delete one of the new entries but first copy its CSV comment into the remaining remaining transaction. So one ledger transaction lists both CSV comments, and subsequent imports from either account are duplicate free.