The short version

The tool reads every text fragment and its position with PDF.js, groups fragments on the same line into rows, clusters their x positions into columns, and writes a genuine .xlsx with each value in its own cell. Purely numeric values are written as numbers so they sum and sort.

Clean, grid-aligned tables come out matching the PDF. Wrapped cells, merged headers and cramped columns can push a value one cell over. Expect a minute of tidying, not a retype. Scans need OCR first.

Bank statements, invoices, price lists, the appendix table in a report: the numbers are right there on the page and completely useless to a spreadsheet until someone types them in. The converter above does the typing. It is a strong first pass rather than a perfect one, and knowing where it is likely to stumble turns the cleanup into a quick scan instead of a hunt.

How to convert a PDF to Excel

1

Confirm the table is text

Select a number in the PDF with your cursor. If it highlights on its own, the page has a text layer. If the whole page selects, it is a scan and needs OCR first.

2

Add the PDF

Drop it or browse. It is read into browser memory; nothing is uploaded.

3

Press Convert to Excel

Text positions are read, rows and columns detected, and the .xlsx assembled in the browser. Each page's table becomes its own worksheet.

4

Open the .xlsx and tidy

Check the header row, scan for shifted cells, confirm numbers are numbers. Then sum, sort and filter as usual.

Output
Real .xlsx

One value per cell, one worksheet per page. Opens in Excel, Sheets and LibreOffice.

Numbers written as
Numbers

Values that are purely numeric can be summed immediately. Anything with a symbol stays text.

Bytes uploaded
0

The tables people most want out of PDFs are financial, and they stay on your machine.

How the detection works

Every piece of text in a PDF comes with coordinates. The converter collects them all, then makes two decisions. Fragments whose vertical position matches, within a tolerance, belong to the same row. Across all rows, the horizontal starting positions cluster into a small number of groups, and each group is a column. A fragment lands in the cell where its row and its column meet.

That is enough to reproduce a table whose columns line up consistently, which describes most exported reports, statements and invoices. It is also why certain layouts confuse it: anything that breaks the assumption that a column is a vertical stripe of aligned text.

Table by tableWhat to expect
  • Exported reports with aligned columnsMatches the PDF
  • Bank and card statementsUsually clean
  • Invoices with line itemsUsually clean
  • Cells wrapping onto two linesSecond line may become its own row
  • Merged or two-line headersHeader may split
  • Dense financial statementsClose columns can merge
  • Scanned tablesNothing to read; OCR first

The two-minute cleanup

Even a good conversion benefits from a glance, and the problems are predictable enough to check in order.

  • Header row. A column title that wrapped in the PDF may have become a second row. Merge the two or delete the fragment.
  • Shifted cells. A value in an unusual position can land one column over. It is obvious at a glance because the column stops lining up. Cut and paste it back.
  • Numbers that are text. A figure with a currency symbol, a trailing CR, or thousands separators in an unexpected style stays text. Strip the symbol, or use Excel's Text to Columns, and it becomes a number.
  • Blank spacer rows. Some layouts leave empty rows where the PDF had visual spacing. Filter and delete.
  • Totals. Recompute a column total and compare it with the PDF's. A match confirms nothing shifted; a mismatch points at the row to inspect.
Statements have their own tool
For a bank or card statement, Bank statement to CSV looks specifically for a date, a description and a signed amount on each line, normalises the dates and signs the amounts, which is usually cleaner than a general table conversion for that one document type.

When to use this, and when not to

Convert to Excel when

You need the numbers as numbers

  • Reusing a report's figures in a chart or a model
  • Bank data into a budget or bookkeeping sheet
  • Invoice line items into accounts
  • Combining tables from several PDFs into one workbook
  • Any table you were about to retype
Use something else when

The table is not the point

Scanned tables and OCR

ScansTwo steps
Scanned tablepixels onlyOCRadds text with positionsPDF to Excelthis tool.xlsxcellsThe converter needs text with coordinates. OCR provides both.
OCR quality decides column alignment. A straight, high-contrast scan gives positions the detector can cluster; a skewed one drifts.

A scanned statement is an image, so there are no text positions to read and the conversion produces nothing. Run it through an OCR tool first; the OCR writes text fragments with coordinates, which is exactly what the detector needs. Straighten and crop the scan before OCR if you can, because a skewed page makes every column drift diagonally and the clusters blur into each other.

Why the tables never leave your browser

The tables people most want out of PDFs are the sensitive ones: statements, payroll, pricing, a household budget. Uploading the file to a conversion server exposes every figure to a third party in exchange for a format change.

Two ways to convertOne of them never transmits a number
Your devicePDFin memoryBrowser enginedetect tablesDownload0 bytes leavethe machineTypical online toolPDFon diskuploadServercopies filedownloadA full copy now sits on hardware you do notcontrol, under a policy you cannot audit
PDF.js extracts the positioned text inside the tab, and the .xlsx is assembled in the browser from the same ZIP-and-XML parts Excel writes. No request leaves your device with the data inside it.

The tool is right above

Everything on this page runs in your browser. Scroll up, drop the file, and keep it on your machine.

Jump to the tool

Common questions

What people ask with the spreadsheet open.

01How accurate is the table detection?

Clean, grid-aligned tables come out matching the PDF. Wrapped cells, merged headers and very close columns can push a value one cell over. Treat it as a strong first pass with a minute of tidying, not a retype.

02Is the spreadsheet editable?

Yes. It is a real .xlsx with each value in its own cell. Open it in Excel, Google Sheets or LibreOffice and work with it normally.

03Does each page become a sheet?

Yes. Each page's detected table becomes its own worksheet, so a multi-page report stays organised and nothing from page 3 lands under page 2.

04Do numbers stay numbers?

Values that are purely numeric are written as numbers so they sum and sort. A figure with a currency symbol or stray character stays text until the symbol is removed.

05Will it work on a scanned PDF?

Only after OCR has added a text layer with positions. A pure image scan has nothing to read. Straighten the scan before OCR for better column alignment.

06What about bank statements specifically?

The statement tool is usually cleaner for those, because it looks for the date-description-amount shape and normalises dates and signs.

07Can I convert a password-protected PDF?

Not while it is encrypted. Remove the password from a file you own with the Unlock tool first, then convert.

08Is it free?

Free. Every tool, unlimited use, no signup.