Extract tables from a PDF into a real .xlsx workbook with automatic row & column detection. Free, in your browser.
🔒 Processed entirely in your browser. Your PDF is never uploaded.
It's extraction with a positional heuristic, packaged into a real .xlsx workbook — not a "true" table-recognition engine. PDFs don't store tables as structured tables internally, only text positioned on a page, so this tool clusters text by vertical position into rows and by horizontal gaps into columns. That's the same honest limitation any free browser-based tool has, and it works well on clean, regularly-spaced tables like most bank statements and simple reports. What is genuine: the output is a real spreadsheet file, generated with the same library this site's other spreadsheet tools use, not a text file wearing an .xlsx extension.
Is this a genuine table conversion, or extraction with guesswork?
It's extraction with a positional heuristic, built into a real .xlsx workbook. PDFs don't store tables as tables internally — just text positioned on a page — so this tool clusters text by vertical position into rows and by horizontal gaps into columns. It works well on clean, regularly-spaced tables and less reliably on merged cells or unusual layouts.
Do I get a real .xlsx file, or just a renamed CSV?
A real .xlsx workbook, built with the same open-source library approach used across the site's spreadsheet tools — it opens directly in Excel, Google Sheets, or LibreOffice Calc as a proper spreadsheet, not a disguised text file.
Why did some columns get merged or split incorrectly?
Column detection relies on the horizontal spacing between words. Unusually tight or wide spacing, or cell text that wraps across multiple lines, can occasionally cause the row/column grouping to merge or split incorrectly. Always check the preview before relying on the output.
Does this work well for bank statements?
Reasonably well for statements with a clear, regular column layout (date, description, amount, balance). Heavily styled or multi-line transaction descriptions can still cause row misalignment — review the output before relying on it.
Does this work on scanned PDFs?
No. It reads the PDF's actual embedded text layer, not pixels — a scanned PDF with no real text layer has no positional data to extract from.
Is my PDF uploaded anywhere?
No. Extraction and spreadsheet generation both happen entirely in your browser — your file is never sent to a server.