Home › PDF check
Is my bank statement PDF text-based or scanned?
Drop a PDF to see whether it has a text layer, how many pages it has and how much text each page holds. It runs in your browser with pdf.js. The file is not uploaded.
Drop a PDF here or
Read on this computer only.
How the check works
For every page we count the characters pdf.js can extract and whether the page draws images. A page with readable text is text-based. A page with images but almost no text is likely a scan. We also count lines that look like a date followed by an amount, a rough sign of a transaction table. Text that's present but invisible (an OCR layer on top of a scan) still counts as text; how well it converts depends on the OCR quality.
Text-based? Convert it to CSV or QBO
StatementHaul converts text-based statements in your browser and checks that opening + credits − debits = closing. 5 statements a month are free, no signup.
FAQ
Is my PDF uploaded?
No. The checker opens the file with pdf.js inside your browser. Nothing is sent to a server; only an anonymous event with the verdict and page count is recorded.
What does text-based mean?
The PDF contains real characters you can select and copy, as statements downloaded from online banking usually do. A scanned or photographed statement is only a picture of the page, with no text layer.
My PDF is scanned. What now?
Download the statement again from online banking instead of scanning a printout, or ask the client or bank for the electronic PDF. StatementHaul does not do OCR, so scanned statements can't be converted.
Does a text-based verdict guarantee the statement converts?
No. It means the text can be read. The converter still has to recognise the transaction layout, and it checks the result against the opening and closing balances. Statements that fail or don't balance are not counted against your free statements.
Why does it say mixed?
Some pages have text and others don't, for example a scanned page added to an electronic statement. Only the pages with text can be read.