Skip to content

PDF to Word

Convert a PDF into a .docx made of real paragraphs and Word heading styles, so you can edit it normally instead of dragging text boxes around.

Private · on device. Runs in your browser: your file is not uploaded. No sign-up.

Nothing is uploaded, it is read on this device
Nothing is uploaded. The PDF is read in this browser tab and never sent anywhere, no account, no page limit, no watermark. You can confirm it by watching the Network tab in your browser's developer tools, or by disconnecting from the internet after this page has loaded.

The text box problem

It is the most common complaint about PDF-to-Word conversion, and it sounds trivial until you meet it. The file opens. It looks right. Then you try to change a sentence and the whole thing falls apart, the text will not reflow, the cursor jumps somewhere unexpected, and deleting a line leaves a hole.

What you have is not a document. It is a page with a box drawn on it for every line of the original, each box containing that line of text and pinned to the coordinate it occupied in the PDF.

Why converters do it

Because it is the path of least resistance, and because it makes the "preserves your formatting perfectly" claim technically true.

A PDF has no paragraphs. It has instructions of the form "draw these characters at 214, 588 in 11pt Times". To reproduce the page exactly, a converter only needs to place each of those fragments in a box at the same position. No interpretation required, and the output is a pixel-accurate match.

To produce something editable, the converter has to work out what the fragments mean: which lines belong to the same paragraph, which are headings, which are list items, and what order any of it should be read in. That is inference, it can be wrong, and it produces a document that no longer matches the original exactly. Most converters decline the problem.

The three inferences that make a document editable

Reading order. A PDF records no order that means anything. On a two-column page the columns share their baselines, so a converter reading across the page interleaves them, a line of the left column, then a line of the right, and produces text nobody can use. The columns have to be found geometrically first.

Paragraph reassembly. This is the one that matters most. A PDF breaks a line wherever the text ran out of column, and those breaks carry no meaning. If they are preserved, every line becomes its own paragraph and the document behaves like a stack of index cards. Joining them back up, while keeping the breaks that are meaningful, between real paragraphs and around list items, is what makes the file behave like a Word document.

The tool above reports how many lines it joined, so the work is visible rather than asserted.

Heading detection. Inferred from font size relative to the body text, with two sanity checks: a heading is short, and it does not end in a full stop. Without those, the first line of any paragraph in a slightly larger face becomes a heading. The result is written as a real Word heading style, so the navigation pane and table of contents work.

What you cannot have

Being plain about the trade, because the marketing around this is not.

You cannot have an exact visual match and an editable document. They are opposites. An exact match is positioned boxes; editable is flowing text. Any converter promising both is giving you the first and calling it the second.

So: images are left out (with a note where each page had them), page numbers and running headers are dropped, and precise spacing is not reproduced. What you get is the content, in the right order, with its structure intact, paragraphs, headings, lists and tables, in a file that behaves normally when you type in it. If what you needed was the layout, you did not want Word, you wanted the PDF.

After converting

Read the preview above against the original before downloading. Ten seconds of checking catches the cases inference gets wrong, and the clean-up guide covers the five steps that turn a rough conversion into a finished document in about ten minutes.

If you are converting to feed the document to an AI tool rather than to edit it, the Markdown converter is the better target, it keeps the same structure at a fraction of the token cost.

Frequently asked questions

Why do other converters give me a file full of text boxes?

Because it is the easy way to make the page look identical. A PDF stores fragments of text at coordinates; putting each one in a box at the same coordinate reproduces the appearance exactly and requires no understanding of the content. The cost is that the result is not a document, typing does not reflow anything, because there are no paragraphs, only boxes. This tool rebuilds the paragraphs instead, which is harder and is the entire point.

Will it look exactly like the original?

No, and it cannot, this is a genuine trade rather than a shortcoming. An exactly reproduced layout is a canvas of positioned boxes, which is precisely the thing that cannot be edited. A flowing document reflows when you type, which means it cannot also be pinned to the original coordinates. You get to have one or the other, and this chooses the editable one, because that is what people converting to Word actually want.

What about images and tables?

Tables whose columns line up are rebuilt as real Word tables, one value per cell, with the header row repeated on each page and amounts right-aligned. A PDF does not record that cells belong to a table, so the grid is worked out from where the text sits; check complex tables against the original. Images are not carried across: a grey note marks each page that had them, and the result lists those pages, so nothing goes missing without a word.

Are my headings preserved?

Headings are detected from font size relative to the body text, then written as real Word heading styles. That means the navigation pane works, a table of contents can be generated, and the outline is correct, rather than the document merely containing some larger text. A heading set at body size gets caught too, if it is short, capitalised and unpunctuated.

Is my document uploaded?

No. It is read in this browser tab and never transmitted. No account, no page limit, no watermark. This matters because the documents people convert are often confidential, contracts, reports before publication, medical letters, internal material, and none of it needs to leave your machine for a format conversion.

It says my PDF has no text. What does that mean?

It is a scan: a photograph of a page rather than a document containing text. There is nothing to convert until the words are recognised. Make it searchable first with the OCR PDF tool, then convert the result. Check it yourself by trying to select a sentence in your PDF reader, if you cannot highlight individual words, the text is not there.

Can it open a password-protected PDF?

Yes. Enter the password when asked and the file is opened in your browser. The password is used locally and never transmitted or stored.

Is there a page limit?

No. Nothing is uploaded, so there is no server cost to meter and no reason to cap it. A very long document uses memory in proportion to its length, so keep the tab in the foreground while it works.