GnuCash is right. It's also why I built my own finance app

It would be nice if financial institutions simply created end-user/app centric APIs for consumption... As it is, the best you can do is something akin to puppeteer with a real browser agent and your authentication credentials and scraping... I mean there can be APIs underneath, but there's not a good open resource/repository that I'm aware of...

Would be nice/interesting to see something like that come about, but It's a bit frustrating. A delegated read-only api key against a specific account shouldn't be that hard to do.

I've slowly built up a set of scripts to do this: too many different methods, but eventually I get csv files for each account, then hledger import - and finally I manually fix up details

Hi Fair challenge, and for a lot of personal finance tooling it's exactly right — and honestly for double-entry too, in two cases: if your finances are simple (one account, a salary), it's overkill; and if you track meticulously but never act on the picture, it's just theater with better-looking books.

Where it earns its keep is when the single-stream "money in / money out" view starts lying to you — several accounts, some debt, reimbursements, an asset or two — because a spending pie chart can't tell you where you actually stand. Double-entry forces the books to balance, which catches errors and gives you a real net-worth position instead of a vibe

came to say this. the fact that I don't have to launch a seperate program to add a transaction or my bank's csv is what finally got me to actually keep track.

Big fan of hedger! It's not easy to start but I think it's one of the best ways to keep track of accounting, PTA files are very convenient.

They leave you with lots of options though, which could be a problem if you're starting into accounting

This is what I use & like as well, but I definitely think there's space for a more GUI-focused option that isn't Quickbooks or Gnucash. It's not a good fit for me, though, I require my important tools to be open source. Closed source software has way too many misaligned incentives for me to use it for anything important.

+1 to that. I used https://ledger-cli.org/ and the accompanying Emacs mode for a long long time. I used small slips of paper to record cash spends during the day and by bank statements for online ones. Then kept track. I kept a "reconciliation" account to catch misses and it was quite low after several years.

I somehow fell out of the habit but I really need to get back to it.

What's the difference betweeen this and either Quicken or Microsoft Money?

Did you write the blog post entirely on your own? It reads like AI slop to me but maybe I'm just getting more sensitive. I can't decide whether we are all starting to mimic the same turns of phrase that LLMs like to use or whether just nobody writes anymore. I suspect the latter.

The bank might not appreciate it, but couldn't you extend with a headless browser to scrape your bank occasionally or even on demand when the program launches?

I've been using hledger and I usually just plug in purchases as I make them, but I do that on purpose because it's like a self-balancing checkbook so I'm always aware of what's going where.

To be clear: as a prototype, it's more than fine. As long as the UI isn't the innovation, make it real first and experiment with it as much as you can. Show it to users, see whether they understand it and see the value. As a domain expert, making your ideas real in any way will make it easier to bring others on board who can create a pleasant experience.

The UI you're showing now looks more like a bag of controls thrown on a page by a backend engineer in 2006 and less like a B2C product in 2026. Various boxes all over the place, weird colors, misalignment, inconsistent spacing, confusing what does what, a random calculator(?) in the middle of the screen, unclear abbreviated labels. It needs so much work that a full rethink might be in order.