PDF to Excel

Each page is examined for table structure rather than being read straight through as text. Ruling lines are used where a table has them, and where it does not, the engine infers columns from how the text aligns down the page. A table that continues across a page break is joined back together when its columns line up. Each detected table becomes a sheet or a block of rows, so the numbers arrive in cells rather than as one long string per line.

Where it stops: This finds tables, not meaning. A layout that only looks like a table to a human, such as figures separated by spacing alone with no consistent alignment, may come out merged or split in the wrong places. Merged cells and nested headers flatten, because a spreadsheet grid cannot hold them as they were drawn. A scanned PDF has no text to read at all: run it through PDF OCR first.