How can I tell if a PDF is scanned?

Try to select a line of text in a PDF viewer. If nothing highlights, the page is an image. If it highlights but pastes as gibberish, the text is unreadable. This checker does the same test on every page at once and also spots scans that have OCR text laid over them, which select normally.

Why does it matter for converting a bank statement?

A text PDF is read exactly. A scan has to go through OCR, and OCR can misread a digit, so a converted scan should be checked against the balances the statement prints before anything is imported.

How do I get a text PDF instead of a scan?

Download the statement from your bank's online banking rather than scanning a paper copy or printing it to PDF. Those downloads are almost always generated as text PDFs; the checker will confirm it.

My statement is password protected. Can it still be checked?

Yes. Enter the password when asked. It is used in your browser to open the file and is not stored or sent anywhere.

What does “Scan + OCR text” mean for accuracy?

The text was produced by recognising the image, so it can contain misread characters even though it looks and searches like text. The sample under the result shows the first text the checker extracted; compare a few numbers with the page.

No. It is opened in your browser with pdf.js, the PDF reader Firefox uses, and never leaves your device.

Bankflow

Free tool

Is this statement a scan, or a real text PDF?

See, page by page, whether a PDF's numbers are real text or only a picture.

Drop a statement PDF hereor browse a file from your device.pdfUp to 100 MB per file · One file at a time

Read in your browser. Never uploaded.

Or try a .

The four kinds of statement PDF

Each page is classified from the text it contains and whether an image covers most of it.

Text

The characters are stored in the file, as they are in a statement downloaded from online banking. A converter reads 1,234.56 as exactly that, with no recognition step to get wrong.

Scan with no text

The page is a photograph or scanner image and holds no characters at all. Selecting text in a PDF viewer selects nothing. Every number has to be recognised from pixels.

Scan with OCR text

A page-size image with invisible recognised text laid on top, which is what a scanner app's “searchable PDF” produces. It copies and searches like a text PDF, but the text is only as good as the recognition was.

Unreadable text

Text is present, but its fonts carry no map from shapes to characters, so copying it gives symbols or boxes. Some statement generators produce this. For extraction it behaves like a scan.

Amounts found

A count of values shaped like money, such as 1,234.56, in the extracted text. A statement page with transactions and no amounts found is a sign the text layer is not what it appears to be.

Who made the file

The producer, creator, and dates recorded in the PDF. A bank's statement system, a scanner app, and a print-to-PDF driver each leave a different name, which often explains the result on its own.

Questions

How can I tell if a PDF is scanned?: Try to select a line of text in a PDF viewer. If nothing highlights, the page is an image. If it highlights but pastes as gibberish, the text is unreadable. This checker does the same test on every page at once and also spots scans that have OCR text laid over them, which select normally.
Why does it matter for converting a bank statement?: A text PDF is read exactly. A scan has to go through OCR, and OCR can misread a digit, so a converted scan should be checked against the balances the statement prints before anything is imported.
How do I get a text PDF instead of a scan?: Download the statement from your bank's online banking rather than scanning a paper copy or printing it to PDF. Those downloads are almost always generated as text PDFs; the checker will confirm it.
My statement is password protected. Can it still be checked?: Yes. Enter the password when asked. It is used in your browser to open the file and is not stored or sent anywhere.
What does “Scan + OCR text” mean for accuracy?: The text was produced by recognising the image, so it can contain misread characters even though it looks and searches like text. The sample under the result shows the first text the checker extracted; compare a few numbers with the page.
Is the PDF uploaded?: No. It is opened in your browser with pdf.js, the PDF reader Firefox uses, and never leaves your device.

More free tools

All free tools

Scanned or not, get the transactions out

Bankflow reads text and scanned statement PDFs alike, and checks every page's transactions against the opening and closing balances printed on it, so an OCR misread shows up instead of slipping through.

Convert a statement free How parsing works

50 free page credits · No card