Will this correctly detect a table if the page also has other text on it?

Yes. The tool looks for runs of at least 3 rows whose columns line up, and writes only those tables to the workbook. Headings, captions, paragraphs and two-column text around a table are left out, not forced into cells.

Are numbers and dates real numbers and dates in Excel?

Yes, by default. A column of amounts, currency, percentages or dates is written as real values with a matching format, so SUM, sorting and filters work. Codes, ZIP codes, phone numbers and numbers with leading zeros stay text on purpose. Dates like 09/02/2026 can be day-first or month-first; the tool works it out from the other dates in the column, and if it cannot it leaves them as text and lets you pick the order. You can also turn conversion off to get plain text.

What if my table continues over several pages?

A table whose columns sit in the same place on the next page is joined into one table, and a repeated heading row is dropped. Choose "One sheet per page" instead if you want the pages kept apart, or "Everything on one sheet" to stack all tables on a single sheet.

What if my PDF's table only has 1 or 2 rows?

A table that small is not reliably told apart from ordinary text, so it may not be found. If no table is found you get a clear message and no file, not an empty workbook.

Will this work on a scanned PDF?

Yes. Pages with no selectable text are read with OCR (pick the language of the scan), and the words are then laid out into tables like any other page. The first scan downloads the language data, so it needs an internet connection and is slower; check numbers read from a scan against the original.

What's the difference between this and Excel to PDF?

This pulls table data out of an existing PDF into an editable spreadsheet. Excel to PDF goes the other direction. Use whichever matches the direction you need.

Are my files uploaded to a server?

No. Table detection, OCR and the spreadsheet are all made in your browser. Your PDF never leaves your device.

Will merged cells or multi-line entries convert correctly?

Not reliably. Detection works best on tables laid out in aligned columns. A cell whose text wraps over several lines may come out as extra rows, and merged cells are not recreated. The preview shows the first rows of every sheet so you can check before downloading.

What should I do if no tables were found?

The PDF may be mostly text (try PDF to Word), or its tables may have fewer than 3 rows or be drawn as a picture. If it is a scan, make sure "Read scanned pages with OCR" is on and the language matches.

How does it decide what counts as a table?

It looks for at least three consecutive rows, each made of two or more separate pieces of text, whose columns line up closely. Ordinary paragraphs, bullet lists and two-column article text do not meet that test, so they are not written to the workbook.

Why does it say no tables were found?

Either the pages have no readable text at all, which usually means a scan with OCR switched off, or the text exists but nothing forms a grid of three or more aligned rows. The message tells you which of the two applies.

Does it work on scanned PDFs?

Yes, when Read scanned pages with OCR is on. Pages without selectable text are read with text recognition first, and the result carries a warning to check figures against the original, since scan quality limits accuracy.

What happens with a table that continues on the next page?

With the join option on, a table on the following page whose columns sit in the same places is merged into one, and a repeated heading row is dropped. The join is not offered in One sheet per page mode.

Home/PDF Tools/PDF to Excel

PDF to Excel

Extract tables from PDF to Excel spreadsheets

Files never storedRuns on your device

Drop a PDF with tablesTap to choose a PDF with tables

or click to browseor tap to choose

Extract table data to Excel — processed entirely in your browser

Large files (over 50 MB) may take longer or run out of browser memory

About PDF to Excel

Finds the tables in a PDF and writes them to an Excel workbook you can actually calculate with. Numbers, currency, percentages and dates are written as real values (not text), so SUM, sorting and filters work; a table that carries on over a page break becomes one table, with the repeated heading row dropped; columns of right-aligned amounts stay in one column. Only the tables are written, not the paragraphs around them. Going the other direction? Excel to PDF converts spreadsheet data into a PDF table.

Watch: PDF to Excel in Action

Watch on YouTube · Subscribe for new tutorials

Will this correctly detect a table if the page also has other text on it?: Yes. The tool looks for runs of at least 3 rows whose columns line up, and writes only those tables to the workbook. Headings, captions, paragraphs and two-column text around a table are left out, not forced into cells.
Are numbers and dates real numbers and dates in Excel?: Yes, by default. A column of amounts, currency, percentages or dates is written as real values with a matching format, so SUM, sorting and filters work. Codes, ZIP codes, phone numbers and numbers with leading zeros stay text on purpose. Dates like 09/02/2026 can be day-first or month-first; the tool works it out from the other dates in the column, and if it cannot it leaves them as text and lets you pick the order. You can also turn conversion off to get plain text.
What if my table continues over several pages?: A table whose columns sit in the same place on the next page is joined into one table, and a repeated heading row is dropped. Choose "One sheet per page" instead if you want the pages kept apart, or "Everything on one sheet" to stack all tables on a single sheet.
What if my PDF's table only has 1 or 2 rows?: A table that small is not reliably told apart from ordinary text, so it may not be found. If no table is found you get a clear message and no file, not an empty workbook.
Will this work on a scanned PDF?: Yes. Pages with no selectable text are read with OCR (pick the language of the scan), and the words are then laid out into tables like any other page. The first scan downloads the language data, so it needs an internet connection and is slower; check numbers read from a scan against the original.
What's the difference between this and Excel to PDF?: This pulls table data out of an existing PDF into an editable spreadsheet. Excel to PDF goes the other direction. Use whichever matches the direction you need.
Are my files uploaded to a server?: No. Table detection, OCR and the spreadsheet are all made in your browser. Your PDF never leaves your device.
Will merged cells or multi-line entries convert correctly?: Not reliably. Detection works best on tables laid out in aligned columns. A cell whose text wraps over several lines may come out as extra rows, and merged cells are not recreated. The preview shows the first rows of every sheet so you can check before downloading.
What should I do if no tables were found?: The PDF may be mostly text (try PDF to Word), or its tables may have fewer than 3 rows or be drawn as a picture. If it is a scan, make sure "Read scanned pages with OCR" is on and the language matches.
How does it decide what counts as a table?: It looks for at least three consecutive rows, each made of two or more separate pieces of text, whose columns line up closely. Ordinary paragraphs, bullet lists and two-column article text do not meet that test, so they are not written to the workbook.
Why does it say no tables were found?: Either the pages have no readable text at all, which usually means a scan with OCR switched off, or the text exists but nothing forms a grid of three or more aligned rows. The message tells you which of the two applies.
Does it work on scanned PDFs?: Yes, when Read scanned pages with OCR is on. Pages without selectable text are read with text recognition first, and the result carries a warning to check figures against the original, since scan quality limits accuracy.
What happens with a table that continues on the next page?: With the join option on, a table on the following page whose columns sit in the same places is merged into one, and a repeated heading row is dropped. The join is not offered in One sheet per page mode.

How to use PDF to Excel

Drop a PDF that contains tables. A file bar shows its name and size, and an Options box appears. Under Sheets choose One sheet per table, One sheet per page or Everything on one sheet. Leave Join a table that continues on the next page ticked when tables run over page breaks; it is unavailable in the per-page layout. Keep Write numbers and dates as real values ticked if you want to sum or sort, and if dates like 09/02/2026 are ambiguous, pick Day first or Month first instead of Work it out from the file. Switch on Read scanned pages with OCR for picture pages and choose their language.

Press Extract Tables. You get a count of tables, rows and sheets, any warnings, a preview of each sheet, and the Download Excel File button.

When to use it

Reach for it with bank or card statements, rate tables, invoices with line items, timetables and report appendices, anywhere rows and columns line up and you want to calculate on them. It is especially useful when one table spans several pages and you would otherwise copy it page by page into a single continuous block of rows.

When not to use it

Do not use it on running prose, forms or brochures, because only tables are written and the paragraphs around them are dropped. For a mostly-text document, PDF to Word or Extract Pages & Text is the better route. A table of only two rows is quicker to retype, and one built from merged cells will probably need hand repair in the spreadsheet.

Tips for best results

Read the preview before downloading: numbers and dates are right-aligned there, so a column that sits left-aligned was kept as text. Identifier-like columns such as codes, ZIP codes, phone numbers and values with leading zeros are left as text on purpose, so 00123 keeps its zeros. When dates look wrong, choose the date order yourself instead of leaving it on automatic. Switching the scanned-page option off makes picture pages get skipped and listed in a warning.

Limitations

A table is recognised only as a run of at least three rows whose columns line up, so two-row tables, merged cells and cells whose text wraps over several lines can split into extra rows or be missed. Paragraphs and two-column text are never treated as tables. Charts and images are not carried over. Scanned pages pass through text recognition first, so figures can be misread, and a page yielding fewer than three words is skipped.

Privacy

Text and its positions are read from the PDF inside your browser and the workbook is assembled there, so the PDF itself is never uploaded. If a page is a scan and the OCR option is on, which it is by default, the recognition engine and language data are downloaded once from jsDelivr's hosting for open-source packages; nothing from your document is sent back in return.