AnyFormat

00 PDF → DOCX

PDF to Word

Editable text back out of a PDF, with a straight account of what that costs.

Drop files or a folder here

or choose from your device · paste with ⌘V / Ctrl+V

.pdf

03Options 4

A range like 1-3, a list like 1,4,9, an open end like 10-, or “all”.

On by default. Turn it off for documents where size varies for other reasons — pull quotes, price lists, forms.

Off by default, because a Word document reflows and the PDF’s page boundaries usually stop meaning anything once it does.

Used to decrypt the file in this tab. It is never transmitted — which is the whole reason not to type a document password into an upload-based converter.

  • No upload The conversion runs in this tab. Your file never travels — not to us, and not to the advertising.
  • No sign-up, no email, no daily cap There is no account system to sign up to.
  • No size limit we invented Only your device's memory — about ~2 GB on a desktop browser, ~400 MB on a phone.
  • Turn your Wi-Fi off and convert anyway The conversion needs nothing but this page. Ads will not load without a connection; your file will still convert. That is the whole claim, and it takes five seconds to check.

Pulls the text out of a PDF and writes a Word document you can edit. Paragraphs are rejoined so they reflow, headings are detected, and bold and italic survive. Read the first section before you rely on the result — this is the conversion where the gap between what people want and what is possible is widest, and pretending otherwise would waste your time.

What a PDF actually contains, and why this is hard

A Word document is a structure: paragraphs, headings, lists, tables, in an order. A PDF is not. A PDF contains instructions to draw glyphs at coordinates — put “T” here, put “h” 4.2 points to the right — and nothing that says those glyphs were a sentence.

So converting one back means inferring the document from the marks: which fragments were a line, which lines were a paragraph, which paragraph was a heading. On a PDF exported cleanly from a word processor those inferences are usually right. On a two-column article, a form, a table-heavy report or a scan they are wrong in ways obvious to you and invisible to any converter.

This is why every PDF-to-Word tool disappoints someone. The difference here is that you are told which case you are in, per file, before you find out the hard way.

Tagged PDFs convert well; the rest is reconstruction

Some PDFs carry a description of their own structure — a tag tree recording that this is a heading and that is a paragraph. Word, LibreOffice, InDesign and most report generators put one in when they export properly.

Every conversion here checks for it and tells you what it found. Tagged is the good case, and the result will usually be close. Not tagged means the structure was inferred from geometry, and the result needs checking.

A scan has neither — it is a picture of a page with no text in it at all — and that is reported as an error rather than an empty document.

What comes across, and what does not

Reconstructed: the text; paragraphs, rejoined so they reflow in Word rather than breaking where the PDF's lines happened to break; headings, inferred from text noticeably larger than the body; bold and italic, taken from the names of the embedded fonts.

Not reconstructed: page layout, columns, tables, images, headers and footers, footnotes, colours, fonts and sizes. A two-column page comes out with the columns run together — recovering the reading order of a two-column page from coordinates alone is not reliably possible, and guessing would produce text that looks right and reads wrong.

The document you get is plain: default font, no styling beyond headings and emphasis. That is deliberate. Inventing formatting on top of a reconstruction would make it look more authoritative than it is.

The thing worth trying first

If the result is wrong in a way that matters, a different converter will almost certainly not fix it — the information is not in the file.

Ask whoever made the PDF for the original. Most PDFs were exported from a document that still exists, and five minutes of asking beats an hour of repairing a reconstruction. This tool is for when that is not an option.

Scanned PDFs and what they really are

There are two entirely different things called a PDF.

A born-digital PDF was made by software — exported from Word, generated by a report tool, printed to PDF. It contains real text with real font information, and it converts cleanly in every direction.

A scanned PDF is a photograph of a piece of paper wrapped in a PDF container. There is no text in it at all — only an image — regardless of how clearly you can read it on screen.

The difference decides what is possible. Rasterising works identically on both. Extracting text works only on the first, and a scan yields nothing at all — which is reported plainly rather than handing back an empty file.

Why this one especially should not be uploaded

The PDFs people want back as editable documents are the ones they need to change: a contract to amend, a tenancy agreement, a report to update, a CV, a legal filing, a medical letter.

Every free PDF-to-Word site takes an upload, and several are frank in their terms about retaining files. Nothing here is transmitted. Turn your Wi-Fi off and convert anyway.

Other names for this

Also searched as “pdf to word”, “pdf to doc”, “pdf to word converter”, “convert pdf to word free”.

Questions

Will the layout be preserved?
No. You get the text as an editable document — paragraphs, headings and emphasis — not a copy of the page. Columns, tables, images and page furniture do not survive, because a PDF does not record them as such.
Why did my table come out as a jumble?
Because a table in a PDF is text at coordinates and some lines, with nothing marking it as a table. Reading it back reliably is not possible from geometry alone. The PDF to Text tool in raw-layout mode keeps the columns lined up, which is often more useful.
My PDF is a scan. What now?
There is no text in it to convert — it is pictures of pages. That needs OCR, which recognises characters in an image and is a genuinely different technology. Rasterise the pages with the PDF to JPG tool and run those through an OCR service.
What does “tagged” mean in the result?
That the PDF carries a description of its own structure, put there when it was exported. Those convert well. An untagged PDF has only marks on a page, and everything about its structure was inferred.
Do bold and italic survive?
Usually. They are taken from the names of the fonts the PDF embeds, which is reliable when the producer named them conventionally and silent when it did not.
Is my password safe to type here?
It is used to decrypt the file inside this tab and is never sent anywhere. That is precisely the argument for this page: uploading a password-protected document and its password to a third-party server is asking them to hold both halves of the lock. Turn your Wi-Fi off before typing it if you want to be certain.
Is my PDF uploaded?
No. The rendering happens in this tab using pdf.js — the same engine Firefox uses to display PDFs. Contracts, medical letters, bank statements and legal filings are the documents this conversion is usually for, and none of them leaves your machine.
Can I convert several PDFs at once?
Yes. Drop as many as you like, or a whole folder. Each produces its own set of images, named after the original.