Extract a table from a PDF

How it works

1. 1Upload the PDF (or drag it onto the page).
2. 2We detect the table and reproduce it verbatim — column titles and every row.
3. 3Preview the extracted table on screen.
4. 4Download it as CSV, Excel, JSON or Markdown.

What you download: The table exactly as printed: the document's own column titles and every column, nothing renamed or dropped. Values stay verbatim — we don't reformat numbers or reinterpret dates behind your back.

Why copying a table from a PDF never works

A PDF has no idea it contains a table — it only knows where each character sits on the page. Select-and-copy flattens that into a single stream of text, which is why pasted tables collapse into one column or shuffle values between rows. A real PDF table extractor works from the glyph coordinates: it finds the column boundaries the document actually drew, assigns every word to the cell whose bounds it falls in, and rebuilds the grid. That's what happens here on every upload.

PDF table to Excel, CSV, JSON or Markdown

Different destinations want different shapes. Excel keeps the table typed and human-readable. CSV feeds imports, scripts and databases. JSON gives developers stable keys per row. Markdown turns the table into text an AI assistant or wiki reads natively. One extraction produces all four — pick the download you need, or take several.

Multi-page tables and repeated headers

Long tables spill across pages, and most tools return them as separate fragments with the header row duplicated in the middle of your data. We skip repeated page headers and merge continuation rows, so a 30-page table arrives as one continuous, clean grid.

What about scanned PDFs?

Digital PDFs with a real text layer give exact results — we read the document's own characters and coordinates, not a picture of them. If you upload a pure scan with no selectable text, we detect it and tell you honestly instead of returning garbage rows.

Questions