PDF to ODS

Convert your PDF files to ODS, right here in your browser.

Free
Private
No limits
No upload

... or drop files here

Files never leave your device

Add a PDF file to convert.

Frequently asked questions

What is actually inside a ODS file?

A .ods file is a ZIP archive whose first entry has to be an uncompressed mimetype marker, followed by the XML that holds the content and its styling.

Are my files uploaded to a server?

No. Your PDF is read and the ODS file is built entirely inside your browser, so nothing ever leaves your device.

Can I convert several files at once?

Yes. Add as many PDF files as you like and download the results one by one, or all together as a ZIP archive.

Will the result look exactly like the original PDF?

The text, headings and table columns are recovered and rebuilt as a ODS file. This is a content-faithful conversion rather than a copy of the original page design, so fonts, spacing and images are not carried over.

What happens with a scanned PDF?

A scanned page is a picture of text rather than text itself, so there is nothing to extract from it. Convert it to an image format instead, or run it through OCR first.

Is there a file size or file count limit?

No. Everything runs on your own device, so the only practical limit is how much memory your browser has available.

Do I need to install anything?

No. There is nothing to install and no account to create — the converter runs in any modern browser, on desktop and mobile alike.

How are columns detected without borders?

By measuring the gaps between text. A word space is about a quarter of the type size, so a gap approaching the full size of the type is deliberate separation rather than a space. The ruled lines of a table are vector graphics stored separately from the text and are no help at all — which is why an unruled table converts exactly as well as a fully bordered one.

Does each page become its own sheet?

Yes, so a multi-page statement gives one tab per page and each corresponds to something you can locate in the original.

Why are my numbers stored as text?

Because interpreting them means guessing at locale conventions, and guessing wrong corrupts data quietly — "1.234" is one thousand two hundred and thirty-four in some places and one point two three four in others. Cells come through as text so nothing is silently reinterpreted; convert the columns you need in Calc, where you can see what is being decided.

Should I convert to ODS or XLSX?

ODS for LibreOffice Calc, Google Sheets and open-format requirements; XLSX for Excel. The extraction behind them is identical, so pick whichever your recipient opens without thinking about it.

Two columns ended up in one cell.

The gap between them was too narrow to distinguish from a word space, which happens with dense tables set in small type. Calc's Text to Columns will split the affected column cleanly if there is a consistent separator inside it.

My spreadsheet is empty.

The PDF has no text layer — it is a scan. Try selecting a sentence in a viewer: if nothing highlights, there is nothing to extract and the file needs OCR first.

ODS is the OpenDocument spreadsheet — what LibreOffice Calc saves natively, and the format to ask for when the file is going somewhere that prefers open standards to Microsoft's.

Getting one out of a PDF means solving the table problem, which is real: a PDF has no idea it contains a table. It records where each glyph was drawn, and the columns exist only because someone aligned the numbers and your eye reads the alignment as structure.

The columns are therefore found by measuring. A word space in normally set text runs to about a quarter of the type size; a gap approaching the full size of the type is deliberate separation, which in a document laid out in columns means a column boundary. Each line is split at those gaps and the pieces become the cells of a row.

How to convert PDF to ODS

Add your PDF
Drop it in or choose it. It is read in the page and never uploaded — which is the property that matters when the document is a statement or an invoice.
Convert
Lines are rebuilt from glyph baselines and split at their column gaps, then written as OpenDocument spreadsheet XML inside a correctly ordered ZIP archive.
Check the columns in Calc
Each page becomes its own sheet. Look down the columns before trusting the figures — gap detection is a measurement, and a tightly set table can put two columns in one cell.

Why the table's borders don't help

It seems as though the ruled lines around a table ought to be the obvious way to find it, and they are not usable at all. In a PDF those lines are vector graphics — drawing instructions for strokes — held completely separately from the text. Nothing connects a line on the page to the numbers it appears to enclose.

This has a pleasant consequence: a table with no visible borders detects exactly as well as one with a full grid, because detection was never looking at the grid. What it needs is space between the columns, which is a property of how the table was set rather than of how it was decorated.

There is also a subtlety in how the space is reported. pdf.js emits the gap between two columns as its own whitespace run, as wide as the gap, so a reader that took the runs at face value would see no gap anywhere. Discarding those blank runs is what leaves a measurable distance between the runs that carry text.

What you get

One sheet per page: Named after the page, so each tab corresponds to something you can find in the original document.
Rows and cells: Each reconstructed line is a row, split into cells at the column gaps.
Everything as text: Cells are written as text rather than typed numbers, deliberately — deciding whether "1.234" means one thousand or one and a bit depends on locale conventions, and guessing silently corrupts data. Convert the columns you need in Calc, where you can see the choice being made.
No formatting or formulas: A PDF contains neither. There are no formulas to recover — they were evaluated long before the PDF was made — and cell styling is not carried across.
Merged cells: Not represented. A header spanning several columns is one run of text and lands in the first cell of its row.

ODS or XLSX?

The reconstruction is the same; only the container differs. Choose ODS for LibreOffice Calc, for Google Sheets, and for any organisation that specifies open formats. Choose XLSX if the file is going to someone working in Excel.

Both are ZIP archives of XML, and both are written here in the browser. The ODS writer has to get one detail right that trips up naive implementations: the `mimetype` entry must be the first file in the archive and stored uncompressed, because ODF readers sniff those opening bytes to identify the format rather than parsing the whole file.

About the ODS format

.ods is the spreadsheet format OASIS introduced in 2005. It is published as an open specification under the name ISO/IEC 26300, so a file can be read without the program that wrote it.

A .ods file is a ZIP archive whose first entry has to be an uncompressed mimetype marker, followed by the XML that holds the content and its styling.

Files like this usually come out of LibreOffice Calc, Apache OpenOffice, and Collabora Online.

What comes through

cell values
column boundaries
table rows and columns
reading order

What does not

formulas
cell formatting
embedded images

Worth knowing

This is the format LibreOffice and Apache OpenOffice save in by default, and the one several governments standardised on precisely because no single vendor controls it.
Column boundaries are worked out from where the text actually sits, so a wide table stays a table instead of collapsing into a run of loose words.
A scanned PDF is a picture of a page, not text, so there is nothing for this tool to extract — the result will be empty. That is the honest outcome, not a failure.
Spreadsheet cells arrive as the values shown on the page. The formulas behind them are not in the PDF to recover.

Information