Tabulyn

Extract tables from a PDF

Find every table in a document and turn it into real rows and columns you can edit, calculate with, and export. Works on digital PDFs — reports, statements, exports from accounting and analytics software.

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

What the extractor handles

Multiple tables per document

Every table is detected and labelled by page and position, so a 12-page report with tables scattered through it comes back organised rather than merged into one blob.

Tables split across pages

When a table continues onto the next page and repeats its header row, the rows are stitched back into one continuous table instead of becoming one sheet per page.

Structure kept rectangular

Merged cells are unmerged, blank spacer rows are dropped when unambiguous, and every row is padded to the same column count — what a spreadsheet actually needs.

Values that stay values

Currency, percentages, whole numbers, decimals, and ISO dates are detected during extraction and written to Excel as typed values, so totals and formulas work immediately.

The one thing that decides whether it works

Open your PDF and try to select a value inside the table with your cursor. If the text highlights, the document is digital and extraction will work. If nothing highlights, the page is an image — a scan or a photo — and the text has to be recognised with OCR first, which Tabulyn doesn't support yet.

This single check predicts the result more reliably than anything else about the file: its size, its page count, or which program produced it.

Where people use this

Finance and accounting

Moving statement and ledger tables into a working spreadsheet.

Research and academia

Pulling results tables out of papers for reanalysis.

Operations

Turning supplier and inventory reports into something sortable.

Data work

Getting a quick structured export without writing a parser.

Questions about table extraction

How does table detection actually work?

A PDF stores characters at fixed coordinates plus the lines drawn around them — it has no concept of a row or a column. The extractor reads those ruling lines to find the grid, falls back to analysing whitespace alignment when a table has no borders, then assigns each piece of text to the cell whose bounds contain it.

Can it find more than one table on the same page?

Yes. Each table on a page is detected separately and labelled by position — page 2, table 1; page 2, table 2 — so you can export them individually or together as separate worksheets.

What if the table has no borders or grid lines?

Borderless tables are detected using column alignment instead of ruling lines. This works well for consistently aligned data and less well when columns are ragged, so borderless results are worth a closer look in the preview before exporting.

How accurate is the extraction?

For clean digital PDFs with ruled tables, structure is usually reproduced exactly. Accuracy drops with unusual layouts: nested tables, cells spanning several columns, or footnote rows mixed into the data. That's precisely why every table opens in an editable preview — you see what was extracted before committing to a download.

Is there a limit on table size?

No limit on rows or columns within a table. The limits are on the document: 10 MB and 30 pages per PDF on the free tier.

Pick your export format