PDF to Excel

Get the rows out of a PDF and back as an editable XLSX workbook.

Secure server processingPDFWorkingNo account needed

What this tool does

PDF to Excel does not read a table, because a PDF does not contain one — it records where each word was printed, not that a group of words formed a grid. So the conversion works backwards from the printed page. Your document is sent to the processing service, pdftotext extracts the text with the original layout preserved, every non-empty line of that extract becomes a row, and the pieces are written as CSV that LibreOffice imports and saves as XLSX.

Cells are cut wherever two or more whitespace characters run together. That single rule is the entire column model, and knowing it tells you in advance which documents will convert well. A statement or price list whose columns sit apart with a comfortable gap comes out clean. A table set tightly enough that one space divides two columns arrives as one merged cell, and a line of prose that happens to contain a double space is split into two cells that were never a column at all. There is no header detection, no per-column type choice and no per-sheet splitting, which is why the tool offers no settings — none of those things exist behind a switch that was left off. The intermediate CSV also reaches LibreOffice with no import options attached, so whether a value arrives as a number, a date or text is its reader's decision rather than yours.

Everything that is not text is left behind. Ruled lines, merged cells, shading, fonts and images are not carried across; what you get is values in a grid. Each document produces one sheet, not a sheet per page or per table. Because the extract runs page by page, running headers, footers and page numbers land as rows in the middle of the data, and a table continued over a page break repeats its heading wherever the page turned. Expect to delete those rows and to re-split at least one column by hand — this gives you the figures, not a finished workbook.

The work runs on the OmniKitly service, since pdftotext and LibreOffice are native binaries rather than anything a browser carries. The file travels there over an encrypted connection and is deleted along with the workbook and the job record when the retention window expires, six hours by default. A single file may be up to 200 MB and fifty may ride in one job, with ninety seconds allowed for the text extraction and five minutes for the conversion — but an hourly data and job allowance sits underneath those numbers, so a batch anywhere near fifty full-sized files is stopped by fair use long before the per-job ceilings come into it. A scan is not caught before the upload, either: the document goes up, extraction is attempted, and only when it comes back empty does the job stop and name the file, because a picture of a table carries no characters to pull out.

How to use it

  1. Add the PDF. One document is the usual case; up to fifty can go in a single job, each within 200 MB, though the hourly allowance stops a batch of large files well short of that. Every file is converted separately and comes back as its own workbook named after the source.
  2. Run the conversion. Nothing to configure — the column rule is fixed, so no controls are shown that the converter could not honour. Signed out you are asked a short verification question first, because converting is one of the operations that always triggers one; an active signed-in account skips it. The stages you then see are Converting and Finalising, with the text extraction folded into the first rather than reported on its own.
  3. Open the workbook and fix the joins. Delete the header and footer rows the page breaks left in the data, then re-split any column the two-space rule merged, using your spreadsheet's text-to-columns on a character or a fixed width.
  4. If the job is refused, run PDF OCR first. A scan carries an image of the page, so extraction finds nothing and the job stops rather than handing you an empty sheet. Convert the searchable PDF that OCR returns.

Processing & privacy

Secure server processing. This operation requires temporary server processing.

Common use cases

Totalling a bank or card statement

Statements are issued as PDFs with generously spaced columns, which is exactly the shape this heuristic handles best. The transactions land as rows you can sum, sort and reconcile.

Recovering figures when the source file is gone

Nobody can find the original workbook, or the supplier will only send the printed version. Converting the PDF table beats retyping several hundred numbers and getting one of them wrong.

Comparing two published price lists

Convert both and put the columns side by side in a spreadsheet, instead of reading two documents in parallel and trusting your eye to catch the differences.

Feeding printed data into a chart or a model

Anything that expects rows and columns — a pivot table, a plotting script, an import into another system — needs the values loose rather than fixed on a page.

Questions

Is my PDF uploaded?
Yes. pdftotext and LibreOffice both run on the OmniKitly processing service, so the document is sent over an encrypted connection, converted there, and deleted with the workbook and the job record once the six-hour retention expires.
Why did two columns end up in one cell?
They were separated by a single space in the original. Cells are cut only where two or more whitespace characters run together, so a tightly set table merges. Splitting that column afterwards in the spreadsheet is normally quicker than reworking the PDF.
Why has my page of text become a hundred rows?
Every non-empty line of the extract becomes a row, whether or not it was ever tabular. Nothing here tells a paragraph from a data row, so a mostly-prose document converts into a sheet of sentences. PDF to Word is the better destination for that material.
Does it work on a scanned document?
No. A scan stores an image, so there are no characters to pull out. The refusal comes after the upload and a failed extraction rather than up front, and the message names PDF OCR. Run that first, then convert the searchable PDF it produces.
My PDF has a password on it — will it convert?
Not while the password is on it. The job fails at the extraction step. The message you see talks about scanning, because that is the usual cause of an empty extract, so if you know the file is not a scan, remove the password with Unlock PDF and convert the result.
Can I choose the columns, the sheets or the header row?
No, and nothing pretends otherwise. The column rule is a fixed heuristic over a text extract rather than a table model, so header detection, column types and a sheet per table are not available to offer.
Does any formatting survive?
None of it. Borders, merged cells, shading, fonts and embedded images are dropped, leaving values in a grid. If you want the page to look as it did rather than to calculate, PDF to JPG keeps the appearance instead.

Related tools

  • Merge PDF — Combine several PDFs into one document in the order you choose.
  • Split PDF — Split one PDF into separate documents by page range.
  • Compress PDF — Reduce PDF file size while preserving the quality that matters.
  • PDF to JPG — Export every page of a PDF as a JPG image.
  • JPG to PDF — Turn one or many JPG photos into a single PDF.
  • Protect PDF — Add a password and set permissions on a PDF.