How it works
A bank statement PDF has no rows and no columns, only characters placed at coordinates. Turning that back into a table is the entire job.
The problem with template parsers
Most statement converters keep a template per bank: this bank puts the date at x=40, the amount at x=380. That works until the bank changes its statement design, or until someone uploads a statement from a credit union nobody templated. The long tail of banks is enormous, and a template library never covers it.
Reading the document instead
This converter reads the statement the way a person does. It uses the headings, the section structure and the position of the numbers to decide what each row means, which is why a withdrawal printed as a positive number under a withdrawals heading still comes out negative, and why a bank with no template still converts.
What happens to your file
The path is short and nothing is kept:
- Your browser uploads the PDF over HTTPS
- The file is held in memory while the transactions are read
- The extracted rows are returned to your browser as CSV
- The file is discarded; it is never written to disk or stored
Verify every conversion
No converter should be trusted blindly, including this one. Add the opening balance to the sum of the amount column. If it equals the closing balance printed on the statement, every transaction was captured. If it does not, something is missing or a sign is wrong, and it is far cheaper to find that now.
What it does not do
It does not connect to your bank, it does not need credentials, and it cannot fetch statements for you. It reads a PDF you already downloaded. If you need statements you no longer have, request them from the bank first.