How to convert your PDF to a DOCX file

Check whether the text in your PDF can be selected: open it in a PDF program and drag the mouse over a few words. If only a frame around the whole page is selected, or nothing at all, it is a scan. In that case, the section on scanned PDFs below will help.
Click “Choose file” or drag the PDF into the drop area. Exactly one file is converted per run.
Click “Convert to DOCX”. Use “Cancel” to stop processing; no file is created in that case.
Under “Result”, read the “Notes on the result”. They tell you, for example, whether images were left out, tables were recreated or two columns were taken over one after the other.
Click “Download” to save the DOCX file.
Open the saved file and check it as described below before you pass it on or delete the PDF.

Your PDF is not changed in the process; the DOCX file is a new, separate file. The “Pages” figure in the result refers to the PDF. In Word, the page count may differ because the text is reflowed. Use “Reset” to clear the tool for the next file.

Editable text instead of a page image

Some conversions simply place each PDF page as an image in a Word document. That looks like the original, but individual words can neither be selected nor changed. This tool produces the opposite: real paragraphs and, for simple numeric tables, real Word table cells. Page images are not inserted.

The reason a new document is created lies in the PDF format. A PDF mainly stores which text is located where on the page, but not which lines form a paragraph or which values form a table row. These relationships have to be worked out anew in every conversion to Word, which is why the appearance can change. Microsoft explains this for Word’s own way of opening PDFs under Opening PDFs in Word.

How to check whether the file is really editable

A glance at the first page is not enough for this. Test the following in the saved file:

Select a word and type over it.
Change a value in a table cell if the file contains a table.
Add a sentence to a paragraph and check whether the following text moves along.
Save, close and reopen the file. Your changes must still be there.

Layout, fonts, tables and images: what is carried over

The text is carried over, not the design of the page. Here is what that means for the individual components of your PDF:

How components of the PDF appear in the DOCX file
In the PDF	In the DOCX file
Paragraphs and lines	New flowing text at roughly the original font size. Line and page breaks are created anew in your word processor; the page count may change.
Fonts	The original fonts are not carried over; the file is set to Arial. If Arial is missing on your device, your word processor shows a substitute font. Bold and italic text are not reliably retained.
Two text columns	Single-column text: first the complete left column, then the right one. Three columns or unclearly arranged columns cause the file to be rejected.
Tables	Simple tables with 2 to 6 columns, at least 3 rows and numeric values become real Word tables. Merged, nested, irregular or text-only tables are not recreated as tables. Their content may appear as flowing text; if the arrangement is unclear, the file is rejected.
Headings, lists, footnotes	Appear as normal text, without Word styles, automatic numbering or real footnotes.
Images, graphics, backgrounds	Are not carried over. The result points this out.
Comments, highlights, links	Are not carried over; links are no longer clickable.
Bookmarks and document properties	Are not carried over.
Blank pages	Are retained as empty sections. The result points this out.

When the appearance has to match exactly

If you need the original page appearance, for example for a letterhead, a flyer or pages with many graphics, converting to new flowing text is the wrong approach. The best option is the original file from which the PDF was created. Ask the sender for it. If you only want to show a page unchanged, you can save it as JPG and insert it as an image, but the text is then not editable. Word can also open PDFs itself; according to Microsoft, the appearance may change in that case too.

Excerpt of the first page of the sample PDF with German text and a table. The content includes “München” and “1.234,56”, among other things. — In the DOCX file, the content is available as editable text and the table as a real table. In this example, umlauts as in “München” and numbers such as “1.234,56” were preserved. However, the lines wrap differently than in the original, and the table was rebuilt with border lines.
The preview comes from LibreOffice, not Microsoft Word. Which font is displayed depends on the fonts on your device. Layout and page breaks may differ from the original.

Excerpt of the first page of the converted DOCX file, displayed in LibreOffice Writer. The text is the same but wraps at different points. The table has been rebuilt and has borders. — In the DOCX file, the content is available as editable text and the table as a real table. In this example, umlauts as in “München” and numbers such as “1.234,56” were preserved. However, the lines wrap differently than in the original, and the table was rebuilt with border lines.
The preview comes from LibreOffice, not Microsoft Word. Which font is displayed depends on the fonts on your device. Layout and page breaks may differ from the original.

Scanned PDF: recognise text first

To a computer, a scan is a photo of the page. It contains no text that could be selected. This tool only reads existing text and does not recognise lettering in images. If even a single page consists solely of an image, the entire conversion therefore stops, even if the other pages contain text. Here is how to proceed:

Make the PDF searchable with text recognition (OCR). It processes up to 20 pages per run and recognises German, English or both.
Check the recognised text, especially names, numbers and amounts. Otherwise, text recognition errors end up unchanged in the DOCX file.
Then select the searchable PDF here and convert it.

This also applies to PDFs that already have a text layer, for example ones coming directly from a scanner. This tool takes over the existing text as it is and does not check whether it was recognised correctly. If the result reports that images were not carried over, check whether any of them contained text. Otherwise, a scanned section on an otherwise digital page will be missing from the Word file. Text recognition does not process such mixed pages either, so you will then have to add that section yourself.

Checking the result: what matters for German documents

For German documents, special characters, number formats and identifiers matter most. In a German sample document, entries such as “Straße 12”, “Äpfel, Öl und süße Grüße”, the amount “1.234,56”, the invoice number “DE-2026-4711” and the easily confused sequence “0 O 1 l I” arrived unchanged in the DOCX file. Whether this also applies to your PDF depends on how its text was stored or recognised. Therefore, compare the result with the original:

Names, streets and places with ä, ö, ü and ß
Amounts with a thousands point and decimal comma, as well as dates
Invoice, customer and file reference numbers in which 0 and O or 1, l and I can be confused
the order of the sections, especially on two-column pages, footnotes and footers
recreated tables row by row: is every value in the correct cell?
content that was present in the PDF as an image, graphic or comment and is therefore missing

If part of the text cannot be read out as legible characters, the conversion stops with a message. Recreating the PDF from the original program usually helps in that case. Once the file has been fully edited and is to become a PDF again, save or export it directly as a PDF in Word or your word processor. Convert Word to PDF is only an option if your document consists of a single Word section. Results from multi-page PDFs contain a separate section for each page; this converter does not accept such files.

Which PDFs can be converted

Per run, you can convert one PDF with up to 50 MiB, 100 pages and 300,000 characters of text. The text must run horizontally and from left to right, as is usual in German documents. The result is always a DOCX file; the older DOC format is not produced.

When conversion is rejected and what you can do
Your PDF …	What you can do
is encrypted or password-protected	Remove the protection, provided you are authorised to do so, and try again. The tool does not bypass any protection.
is digitally signed	Use a copy without a signature. The signature is neither removed nor verified here.
contains form fields, file attachments or layers	Use a version without these components.
has pages that consist only of an image	Use text recognition first, as described above.
has rotated pages or vertical or right-to-left text	Use a version with normally oriented pages and horizontal text.
is larger than 50 MiB, has more than 100 pages, more than 300,000 characters or very many text elements on one page	Split the PDF into smaller parts that each lie within the limits stated above, and convert the parts individually. At no more than 50 MiB and 250 pages, you can split it into smaller parts here. If it is larger or longer, recreate it in smaller parts in the original program or split it with a PDF program on your computer. If a single page contains too much text, that is, more than 20,000 individual text pieces, splitting does not help, because that page remains just as dense. In that case, export the PDF again from the original program or simplify the page there, for example by spreading its content over several pages.
has more than two or unclearly arranged columns, overlapping text or a very unusual page format	Use a version with a simpler layout. The width and height of every page must be between about 5 and 56 centimetres, so A4 and A3 fit.

You don’t need an account for the conversion, there is no payment step in the tool, and you don’t install anything; the conversion runs in a current browser. You can find the currently applicable conditions under Prices and limits.

Does the PDF stay on your device?

For the conversion itself, yes. The PDF is read in your browser’s memory, and the DOCX file is also created there. The document is not sent to a server for this. However, your browser loads the web page and the tool’s program components from this website in the normal way. You therefore need an internet connection when you open the page; conversion entirely without a network connection is not guaranteed.

Via “Download”, the DOCX file is saved wherever your browser stores downloads. If you are working on a shared device, delete the copy after use. The Security page explains how file processing works on this website. Information on how data is handled can be found in the privacy policy.

Convert PDF to Word

Result

Notes on the result