PDF to Excel

Pull tables out of a digitally generated PDF into a real .xlsx workbook you can edit.

This reads the text layer a PDF carries, so it works on PDFs produced by software — exports, statements, reports. It cannot read scanned pages or tables saved as pictures, because there is no text in them to read.

Drop a PDF here or browse

PDF · max 50 MB · up to 100 pages

What this tool does

It reads the text layer inside a PDF, works out which parts of each page are laid out as a grid, and writes each table to its own sheet in an Excel workbook. Values that are genuinely numeric arrive as numbers, so they can be summed straight away.

A table that continues across a page break is joined back together, and a repeated header row on the second page is dropped. Only tables on consecutive pages with the same number of columns are joined, so two unrelated tables are never silently merged.

Important limitations

This works on PDFs generated by software — exports, statements, system reports. It reads the text a PDF already carries.

It cannot read scanned documents or tables saved as pictures, because there is no text in them to read. If a document has no text layer, you are told that rather than being given an empty workbook.

Complex magazine-style layouts, tables without consistent column positions, and heavily merged cells will not extract reliably. Where nothing table-shaped is found, nothing is invented.

When you might need it

You need to do arithmetic on data stuck in a PDF: Totals, sorting and filtering are all impossible until the numbers are in a spreadsheet.
Retyping would take an hour: The usual alternative to this tool, and the one that introduces mistakes.
A system only exports PDF: Plenty of older business software offers a printable report and nothing else.

Common workplace uses

Bank and card statements: Get transactions into a sheet for reconciliation or expense categorisation.
Supplier invoices and purchase orders: Pull line items out to compare against what was ordered or received.
Reports from internal systems: Turn a fixed PDF report into something you can pivot and chart.
Price lists and catalogues: Bring a supplier’s list into a sheet to compare against your own.

How to use it

Add your PDF
Up to 50 MB and 100 pages. Each page is analysed as it loads.
Look at what was found
Every detected table is previewed with its page number, row and column counts, and the first few rows.
Decide about joining
Tables continuing across pages are joined by default. Turn it off if you would rather each page stayed separate.
Check the preview before downloading
This is the moment to spot a table that has been read badly — much easier here than after opening the workbook.
Download the .xlsx
Each table becomes its own sheet, named with the pages it came from.

Practical tips

Test whether your PDF has a text layer: Open it in a reader and try to select a word inside the table. If you cannot, the page is an image and this tool cannot read it.
Always check the preview: Table extraction is inference, not a guaranteed conversion. A misread column is obvious in the preview and easy to miss in a workbook.
Identifiers stay text on purpose: Values like 007 or a reference number are left as text, because converting them would lose the leading zero. Only unambiguous numbers become numbers.
Ruled lines do not help: Detection works from where the text sits, not from drawn borders. A table with no lines can extract perfectly; one with heavy borders can still be missed if its columns do not line up.
For a scan, OCR first: Running OCR adds a text layer that this tool can then read — but OCR output is a best guess, so check the numbers carefully.

Common questions

Why did it find no tables in my PDF?

Either the document has no text layer, which means it is a scan or images and there is nothing to read, or its content is not laid out as a consistent grid. You are told which of the two it is. Nothing is guessed at, because an invented table is worse than none.

Does it work on scanned documents?

No. A scan is a picture of a page with no text inside it. Run it through OCR PDF first to add a text layer, then try again — and check the results, because OCR makes mistakes.

Is the output a real Excel file?

Yes, a genuine .xlsx workbook. It opens in Excel, LibreOffice, Numbers and Google Sheets, with numbers stored as numbers rather than text.

Why is one table split across two sheets?

Joining only happens for tables on consecutive pages with the same number of columns. If a column was misread on one page, the shapes differ and the tables stay separate — a deliberately cautious rule, because wrongly merging two tables is harder to notice than a split one.

Are formatting, colours and merged cells kept?

No. This extracts values, not appearance. You get the data with a bold header row and nothing else.

Is my document uploaded?

No. The PDF is read and the workbook is built in your browser, and nothing is stored after you close the tab.

Related workflows

Get data out of a scanned report

When the PDF is an image, add a text layer first.

🔍 OCR PDFAdd a text layer with OCR
📊 PDF to ExcelExtract the tables
Check the numbers against the original

Pull one table out of a long report

Narrow the document down before extracting.

✂️ Split PDFKeep only the pages with the table
📊 PDF to ExcelConvert those pages