About Bank Extract

Manual data entry from bank PDFs is one of those tasks that's so painfully boring, so obviously automatable, that surely someone had already fixed it.

I looked. What I found was a graveyard of OCR tools that guessed at numbers, AI wrappers that hallucinated decimals, and cloud services that asked you to upload your bank statements to someone else's server. For accountants — people whose entire job depends on exact numbers — none of that was acceptable.

Your bank statements contain some of the most sensitive data you own. Account numbers. Transaction histories. Client names.

That data should never touch a server. Not mine. Not anyone's.

Bank Extract processes everything locally on your machine. The web app runs entirely in your browser. I physically cannot see your financial data, and I prefer it that way.

Bank Extract doesn't rely on bank-specific templates or pre-configured profiles. It reads the raw coordinate data embedded in any PDF that contains a table. That means it works with virtually every bank in the world.

The key is the column selection: once you draw your columns on the canvas, the extractor maps every row. If rows come out merged or misaligned, the built-in row separator and filter tools let you fix it without re-processing every single time.

Some PDFs have unusual formatting — extra whitespace, merged cells, or inconsistent spacing. The editor includes filters and a find-and-replace tool specifically for these cases. And once you save a template for a particular bank, you never have to reconfigure it again.