How to convert PDF tables to Excel
Copying a table out of a PDF by hand usually breaks it — columns collapse into one, numbers arrive as text, and totals stop adding up. This page explains how to get a spreadsheet you can actually calculate with, and what to do when a table comes out wrong.
Drop your PDF here, or browse files
PDF · up to 10 MB · up to 30 pages
- No signup needed
- Auto table detection
- Files auto-deleted
Why copy and paste breaks PDF tables
A PDF doesn't store a table as rows and columns. It stores characters at fixed coordinates on a page, plus lines drawn around them. The grid you see is visual, not structural — which is why pasting into Excel dumps everything into a single column. Extracting a table means reconstructing that structure from position data, then deciding which text belongs in which cell.
That reconstruction is a best guess, and it is occasionally wrong. This is why Tabulyn shows you every detected table in an editable preview before you download anything, rather than handing you a file and hoping it's right.
Keeping numbers as numbers
The most common complaint about PDF-to-Excel tools is that the output looks correct but behaves like text. Tabulyn classifies each cell as it extracts and writes real typed values into the worksheet:
| In the PDF | In your spreadsheet |
|---|---|
| $1,500.00 | Currency value you can sum |
| 12% | Percentage value |
| 1,500 | Number, not text |
| 2026-08-20 | Date |
| Blank spacer row | Removed when unambiguous |
Anything ambiguous is left exactly as it appeared. The tool won't invent a value or reformat something it isn't sure about — a wrong guess is worse than an honest one you can correct in the preview.
What works, and what doesn't
✓ Works well
- Financial statements and accounting exports
- Invoices, purchase orders, and receipts with line items
- Research papers and reports with data tables
- Bank statements exported from online banking
- Any PDF where you can select the text with your cursor
! Not supported yet
- Scanned documents and photographed pages (needs OCR — coming later)
- Handwritten tables
- Password-protected PDFs (remove the password first)
- Files over 10 MB or 30 pages on the free tier
If a table comes out wrong
- Columns are merged together. Usually means the table has no ruled lines. Check the preview and split the values manually, or try a version of the document exported directly from the source software.
- Only part of the table appeared. If the table continues across pages without repeating its header row, each page is treated as its own table. Select both in the preview — they export as separate sheets you can stack.
- Nothing was detected. Try selecting text in the PDF with your cursor. If you can't, it's a scan, and OCR isn't supported yet.
Questions about Excel export
Why do my numbers come into Excel as text?
Most converters paste every cell as a string, so SUM and AVERAGE silently return zero. Tabulyn detects currency, percentages, integers, decimals, and ISO dates during extraction and writes them to the worksheet as real typed values, so formulas work on the file you download.
Can it handle a table that spans several pages?
Yes. When a table continues onto the next page and repeats its header row, Tabulyn recognises the repeated header and merges the rows back into one continuous table, so you get a single worksheet instead of one sheet per page.
What happens with merged cells?
Merged cells are unmerged into a rectangular grid, which is what a spreadsheet needs. The value lands in the first cell of the span and the remaining cells are left empty rather than duplicated, so you can decide how to fill them.
Will it work on a scanned PDF or a photo of a document?
Not yet. Tabulyn currently reads digital PDFs — files where you can select the text with your cursor. A scan is an image, which needs OCR; that is planned but not available today. If you can't select text in your PDF, it won't extract.
Do I get one worksheet per table?
Yes. Each table you select becomes its own worksheet, named by where it came from — P2_Table1 means page 2, table 1 — so a multi-table report stays organised instead of being flattened together.
Need CSV instead?
The same extraction engine exports plain CSV — useful for scripts, databases, and imports.
Convert PDF to CSV