Hey Indie Hackers,
Quick update on Bank To Ledger (https://banktoledger.com). Following up on my last post, I've spent the week hardening our extraction accuracy for accountants and bookkeepers dealing with shifted columns.
I just deployed a major feature: Automated Running Balance Validation.
Now, the parser doesn't just scrape text—the code actively recalculates the math on every single transaction row against the bank's final statement balance. If a digit is misread or a column shifts, the engine catches the anomaly instantly before exporting.
A reminder on privacy since a few people messaged me about it: everything runs strictly in-memory (RAM). There are no hard disk writes, and we’ve added an instant "Manual Delete" button so you can wipe your data the second your download finishes.
The engine natively spits out clean XLSX, CSV, XML, and OFX files.
I really want to put this validation logic through its paces. If you handle client books, go run a messy or scanned PDF statement through it.
Did the balance validation catch everything? What specific regional bank layout should I optimize the regex for next? Drop your feedback below!
try it here : https://banktoledger.com/