Thanks for the link to weboob project, I haven't seen that one. boobank looks much more like something I want to have, though I can't find anything on how it works - how are the banking modules implemented? In particular, are they accessing an API, or just scrapping the bank's page? The reason myself and others are pushing for consumer-open API access is because, while scrapping is something a competent programmer can do (+/- jumping through the increasingly arcane authorization hoops), it's technically against ToS, and I'm personally not willing to risk my bank thinking I'm trying to "hack" them.
Weboob relies entirely on web scraping: like I said, it's a form of browsing the site, just not with the usual browser. I do agree that services should provide APIs, but webood is targeted at websites won't/don't want to do it.
weboob has many capabilities, such as "listing messages" or "listening to music" and an important one is "getting financial information out of an account". Each capability is in the middle of 2 things:
- Applications, which want to implement a given capability; a basic one is boobank (https://weboob.org/applications/boobank) but you can write anything
- Modules, listed in the same page; a module is a website
The weboob project is completely open source and written in Python. If you use the boobank application you can use it to download and store all your financial transactions, and automate it as needed. As the example on the homepage shows, you can pipe it into munin (or any other graphing tool) and see the evolution in your accounts automatically.