PDF to Excel Converter
Convert any PDF table into a real Excel workbook — separate columns, numeric cells, one row per record. Runs inside your browser, so confidential reports never leave the machine they are on.
What "PDF to Excel" is actually asking for
A PDF has no notion of a row or a cell. It is a list of drawing instructions: put these glyphs at these coordinates, draw a line here. When you see a table, you are reading structure that the original software knew about and the PDF format threw away.
Most converters handle this one of two ways. They either run the page as an image through OCR, which invents characters from pixels and mangles small print, or they concatenate text line by line, which produces a spreadsheet with one column and everything crammed into it.
This converter does neither. It reads the exact position of every text run on the page, groups runs that share a baseline into rows, and then clusters their left edges into columns. Because the coordinates are exact — to a tenth of a point — the columns line up even when the source document used proportional fonts and right-aligned money.
Built for the awkward cases
- Repeated headers. A statement that prints the column header on all twelve pages yields one header row, not twelve.
- Wrapped descriptions. A merchant name that spills onto a second line is joined back to its transaction instead of becoming an orphan row.
- Right-aligned money. Amounts are matched to their column by where they start, not where they end, so a short value like
9.00does not drift into the neighbouring column. - Negative formats.
(1,240.00),-1,240.00and1,240.00-all become-1240. - European decimals.
1.234,56is read as one thousand two hundred thirty-four and change, not as a malformed float. - Letterheads and footers. The page furniture is excluded by default, so row one is your actual column header.
When the columns come out wrong
No converter gets every layout right, which is why this one gives you two ways out. The first is the Column detection setting: More columns splits aggressively when two fields are jammed together, Fewer columns merges when a single field got shredded into three. The preview updates as you change it.
The second is the preview itself. Every cell is editable. If one date is malformed or one description lost a character, fix it there before downloading — it takes less time than re-exporting and correcting in Excel.
A note on dates. Dates are exported as the text the PDF contained. That is
deliberate: a statement that prints 02 Sep without a year cannot be turned into a
real date without guessing, and a wrong guess is worse than plain text. Sort still works,
because the format is consistent within a document.
FAQ
Does this work with tables that have no visible borders?
Yes, and it is the more reliable case. Borders are graphics; this converter never looks at them. It only uses where the text sits, so a table drawn with whitespace alone converts exactly as well as one with grid lines.
Can it handle a PDF with several tables on one page?
It exports the main table — the one with the most numeric rows — and skips surrounding prose. If a page contains two unrelated tables, convert it twice with different column settings, or turn off Only export the table to get everything as text rows.
What happens to merged header cells spanning two columns?
The spanning label sits above the columns it covers. Because it has only one start position, it lands in the column where it begins, and the data rows below it keep their own alignment. It is a readable result rather than a perfect one; Excel's merge button is two clicks away.
Is there a page limit?
No. A 300-page PDF converts in one pass. Large files take longer to read, but nothing is truncated and no paywall appears at page 11.
Do you keep a copy of the workbook?
No copy exists to keep. The workbook is assembled in memory and handed to your browser's download mechanism. Closing the tab discards everything.