I'll give the opposing opinion that YNAB's auto-import works well for me (I use it with a local credit union, 2 credit cards, and an online savings account). In fact, I used my own homebrew budgeting software for over a year before I found that YNAB exists and does everything I needed from my own software plus imports from banks automatically which was my biggest pain point with my own solution (it involved painful csv exports every month). I have been happy with YNAB for a year and a half at this point. Your luck may depend on which bank(s) you want it to work with. I believe their bank import api provider was recently acquired by Plaid so they are now using Plaid (as I understand it).
They are using Plaid, and yeah, it varies wildly by bank.
FWIW, I don't know that having a bank that doesn't play nice with Plaid is a deal-killer. There is a school of thought among many YNAB users that you should not use the auto-import, or even csv/ofx/whatever import, because manually entering transactions takes only a few minutes a day, and gives you much better awareness of your outflows. I can see some merit to that, insofar as tightly controlling your spending is kind of YNAB's whole thing.
Even with auto-import, you still have to "approve" each individual transaction that is imported, at which time you can assign it to a category (or approve the auto-suggested category). So that works for me in terms of being aware of spending on a regular basis. And if you manually enter some transactions as you spend, it automatically matches them up with your bank's version of those transactions as you import so you don't end up with duplicates. So it's okay to use it in a hybrid mode where you sometimes enter transactions manually and sometimes wait for them to show up automatically.