How it finds the table
A PDF doesn’t store tables. It stores each piece of text at a position on the page, which is why copying a table from a PDF usually gives one jumbled line. The converter rebuilds the table the way your eye does:
- Lines: text sitting on the same baseline becomes one line.
- Cells: a gap wider than a normal space splits a line into cells.
- Columns: the horizontal ranges that cells share across many lines become columns. Right-aligned amounts and left-aligned text are both handled.
- Rows: every line becomes a row, and wrapped descriptions are joined to the row they belong to.
Columns are worked out across all the pages you convert, so a statement that runs over several pages gives one continuous table with the same columns throughout.
How to use it
- Optionally type which pages to convert, for example
2-5. Leave it empty for all pages. - Pick the CSV separator: comma for most software, semicolon for Excel in much of Europe.
- Choose your PDF. The note shows how many rows and columns were found.
Getting the best result
- Use the original PDF. Download the statement or invoice from the website or app that produced it rather than a scan or a photo.
- Skip pages without tables. Cover pages and terms and conditions only add rows to delete later.
- Check the totals. For financial data, compare the CSV’s sum of a column with the total printed on the PDF. It takes a minute and catches any row that landed in the wrong place.
Layouts with tables side by side, or text flowing around a table, can mix columns together. Most single-table documents, which includes nearly all bank and card statements, convert cleanly.
Privacy
Bank statements are among the most sensitive files people convert. Most PDF converters upload them to a server. Here the PDF is read by pdf.js, Mozilla’s PDF engine, inside your browser, and the CSV is created on your device. Nothing is uploaded or stored.
Questions
Which PDFs work?
PDFs created by software, such as statements downloaded from online banking, invoices from accounting tools or reports exported from other programs. Scanned paper documents are pictures of text and need text recognition (OCR) first, which this tool doesn't do.
Why are some rows that aren't part of the table in my CSV?
The converter keeps all the text on the page, including headings, addresses and footnotes, each on its own row. That way nothing is silently lost. Delete the rows you don't need once the file is open in your spreadsheet.
A long description split over two lines. Is it two rows?
Usually not. When a line holds only text and sits directly under a full row, it's joined to the row above, which keeps wrapped transaction descriptions together.
Should I use CSV or Excel output?
CSV for importing into accounting software, budgeting apps or scripts. Excel if you want to work with the numbers yourself; the Excel version stores amounts as real numbers you can total.
Last reviewed 3 November 2026