PDF to CSV Converter

Bank statements, price lists and reports are often shared as PDFs, but a PDF has no real rows or columns — only words placed at x and y positions. This converter reads those positions, groups words into lines, works out where the columns are from the empty gaps that run down the page, and writes the result as a CSV file that opens in Excel, Google Sheets, Numbers or any database tool. Everything runs in your browser.

How to use this tool

1. Drop one or more text-based PDF files.
2. Optionally limit the pages, e.g. 2-4 for the table pages only.
3. Pick the delimiter: comma for most programs, semicolon for Excel set to a European language, or tab.
4. Tick "Clean up numbers" if you want amounts like 1,234.50 or (12.00) turned into plain numbers.
5. Download the CSV and check the preview table.

Formula

Same line: |y₁ − y₂| ≤ 0.45 × font size. New cell: horizontal gap > 0.6 × font size (Normal). Column boundary: an x position covered by no more than 5% of multi-cell lines.

Lines with two or more cells vote on where the columns are; single-cell lines (titles, notes) are placed in the column where they start. "Only wide gaps" needs a gap of 1.5 × the font size, which keeps phrases with wide word spacing together; "Tight" splits at 0.3 × the font size for densely packed tables.

Example

A statement line with "03/14", "Coffee shop" and "(4.50)" at x = 40, 120 and 480 pt, and gaps of more than one font size between them, becomes 03/14,Coffee shop,(4.50). With "Clean up numbers" on, the amount becomes -4.50, so a spreadsheet can add it up. Lines whose text sits in the same x ranges line up under the same columns.

What it cannot do

Scanned PDFs contain pictures of text, not text, so they produce an empty CSV until you run OCR on them. Tables drawn with merged cells, wrapped text inside a cell or columns that touch each other may need a quick clean-up in the spreadsheet: a wrapped cell becomes two rows, and two touching columns can merge into one.