Four methods · What each really does · Sourced
How to Convert a PDF to an Editable Word Document
How do I convert a PDF to Word? There are four common answers, and they give very different results because a PDF records where each glyph was printed and almost nothing about which glyphs form a paragraph. Making it editable means rebuilding structure the file never stored. Here is how each method rebuilds it, in its own documentation's words.
Upload a PDF Basic mode is free and needs no account: 10 pages a day and 100 a month, rising to 100 pages a day and 1,000 a month once you sign in. Scanned pages come back as page images, with any text the PDF already carries left editable. AI mode transcribes scans into real editable text — sign in and add a payment method, or paste your own API key below. No watermark.
Converting…
Off by default. Your PDF is deleted on schedule unless you tick this; ticked, it is kept for our engineers to study this conversion. Untick any time.
Options · AI / Basic · your own key
Advanced — bring your own AI provider
Scanned pages are read by an AI vision model. Bring your own provider to convert scanned pages at your provider’s rates and the quality you choose.
Applies to your own key. Scanned pages use AI vision either way.
Used only for this one conversion. Never stored, never logged.
Last updated: 2026-09-16
First, thirty seconds of triage. Open the PDF and try to select and copy a sentence. If the words paste out intact, the file is born-digital and every method below has something to work with. If nothing selects, it is a scan, and only the methods that recognise characters will help. Most of the frustration here is a method pointed at the wrong kind of file.
Method 1: open the PDF in Word itself
Word has converted PDFs on its own since PDF Reflow arrived with Word 2013, and Microsoft's current support page lists the feature for Word for Microsoft 365, Word 2024, Word 2021, Word 2019 and Word 2016. Most people who need it already have it.
- In Word, go to File, then Open, and browse to the PDF.
- Word makes a copy and converts its contents; the original PDF is unchanged.
- Edit, then save as .docx.
Microsoft is unusually candid about the limits. Converting from PDF to Word, its documentation says, works best with files that are mostly text; if the PDF contains mostly charts or other graphics, the whole page might show up as an image. It warns that the converted document might not have a perfect page-to-page correspondence with the original, and names tables with cell spacing, frames, footnotes, tracked changes and certain font effects among the content that converts poorly. On copier output it is blunt: something that looks like a manuscript from a copier will not be formatted well in Word.
On a Mac the route is the same File, then Open. Microsoft's page documents the feature by Word version with no separate note for Word for Mac, so the caveats above are the only documented ones.
Use it when: the PDF is short, mostly text, and Word is already open.
Method 2: Google Docs
Google Drive opens a PDF as a Google Doc, running character recognition on the way in when the content is an image. Upload to Drive, right-click, Open with, Google Docs; download as .docx afterwards. It is the only method here that reads a scan with no payment step at all.
Google's help page is specific. The file should be 2 MB or smaller, text at least 10 pixels high, the document right-side up, the image sharp and evenly lit. On output it states plainly that bold, italics, font size, font type and line breaks are likely to be retained — and that lists, tables, columns, footnotes and endnotes are not likely to be detected. Language is detected for you, across more than a hundred including Chinese.
It is also how you convert a PDF to a Google Doc rather than to Word: same action, and the result stays in Drive until File, then Download, then Microsoft Word (.docx).
Use it when: you want the words back from a short scan and the layout is not the point. That 2 MB ceiling is the practical boundary.
Method 3: Acrobat and other paid desktop tools
Adobe invented the format, and Acrobat exports to Word as a first-class feature; the wider category of paid desktop suites works the same way and bundles the neighbouring jobs — redaction, forms, batches. If you handle PDFs daily, it is built for you.
Two things from Adobe's own pages matter. Its free browser converter is described as needing no sign-up, credit card or subscription, while stating that daily conversion limits apply and that unlimited conversions come with a paid Acrobat plan. And Acrobat's help says to set the Document Language to match the text before exporting a scanned document or image — a detail that matters enormously outside English.
No price appears here: plans on tools this size change several times a year. Adobe's pricing page is linked below.
Use it when: PDFs are part of the job.
Method 4: an online converter
The fourth route is a browser: upload, the server converts, you download a .docx. Three questions separate these converters. Does it read the PDF's own text layer, or re-read the page as an image? Does it rebuild real Word objects, or wrap the page in a picture? And what happens to your file after you close the tab?
This site is one of them, built around the last two. You can convert PDF to editable Word here without an account: text pages are rebuilt from the PDF's own geometry, and scanned pages go to an AI vision model that returns structure — headings, paragraphs, table rows — not a flat image.
- Upload the PDF — text-based or scanned, no account needed.
- Each page is classified as text or scan; scanned pages are read by an AI vision model and rebuilt as Word objects.
- Download the .docx and edit it in Word.
There is no watermark on the result. Basic mode is free and needs no account: 10 pages a day and 100 a month, rising to 100 pages a day and 1,000 a month once you sign in. Scanned pages come back as page images, with any text the PDF already carries left editable. AI mode transcribes scans into real editable text — sign in and add a payment method, or paste your own Anthropic, OpenAI, Gemini or OpenAI-compatible API key and pay your provider directly. AI mode, which transcribes scanned pages into real editable text, requires a signed-in account with a payment method on file and draws on that same account allowance; beyond it, purchased pages are charged. Converting with your own API key is not metered at all. Page packs are listed on the pricing page and are not switched on yet. Files are deleted about six hours later. A PDF whose pages are shaped like slides routes to PowerPoint automatically, and the same upload can come back as .xlsx or Markdown.
One thing it is not: a converter that runs with no connection. If you want an offline desktop tool, Word's own import is the free one you already have. A desktop build of ours is on the roadmap, not available today, and gets no date here.
Proof
What editable is supposed to mean
Can ChatGPT or AI do it?
The mechanism, rather than a verdict about any particular assistant.
Two jobs hide inside "convert this PDF". One is reading the page: which marks are characters, which characters form a line, which lines form a cell. The other is writing the document: emitting a file whose paragraphs and table rows behave like Word objects when you edit them. A language model produces text — text it has generated, which makes it excellent at summarising a report and the wrong tool for a contract you must quote word for word. Producing the .docx is a conversion step either way.
Which is not an argument against AI — the opposite. Vision models are the best thing anyone has for reading a scanned page including its structure, which is why this converter uses one. What matters is where the model sits: the reading stage of a pipeline that then writes a real Word file, not the file's author.
The part every method gets wrong: formatting
All four methods above hit the same wall, and it is not a difference in effort: a PDF stores glyph positions, not paragraphs, so every converter has to infer the headings, tables and columns back — and a converter that aims at reproducing the page's appearance will reach for absolutely positioned text boxes, which look right on open and come apart on the first edit. Which of the two you were handed is something you can check in about ten seconds.
The mechanisms, the checks and what we do about each of them are worked through in convert PDF to Word without losing formatting.
When the PDF is a scan
A scan reorders everything above. There is no text layer, so nothing can reflow; the page must be recognised before any of it is rebuilt. Word's importer works from a text layer, which is why Microsoft's documentation points at copier output as poorly served. Google Docs recognises a scan, within its documented size limit. Acrobat handles scans in its paid feature set.
Here, scanned pages go to an AI vision model that reads the page and its structure together, and the result is rebuilt as Word objects rather than embedded as a picture. That is AI mode: it costs real money per page, so it needs you to sign in and add a payment method, with your own API key as the escape hatch.
Scanned source? See scanned PDF to Word.
When the PDF is in Chinese
Chinese and other CJK documents concentrate every difficulty on this page into one file. The text layer is more often unusable, because many Chinese PDFs embed only the glyphs a page draws and omit the table mapping them back to Unicode. Recognition is language-sensitive, which is why Acrobat's help asks you to set the Document Language before exporting a scan. And a .docx stores a typeface's name, not the typeface, so a legacy name that ships with neither Windows nor Office for Mac can turn a page into empty boxes.
All three have concrete fixes, worked through on our page about converting a PDF with Chinese characters to Word.
So which method should you use?
Short text PDF with Word already open: use Word. Small scan where you only want the words: Google Docs. PDFs every day plus redaction, forms and batches: a paid desktop suite. Anything that must stay editable after you change it is where a converter that rebuilds Word objects earns the upload — you can convert PDF to editable Word here with no account, card or install.
Sources
Third-party behaviour above comes from each vendor's own documentation, checked 2026-09-14. Features, limits and prices change; follow the links.
FAQ
Converting PDF to editable Word — questions
How do I convert a PDF to Word on a Mac?
Every method on this page is reachable from a Mac. In Word the route is the same File, then Open, then the PDF; Microsoft documents the feature by Word version and publishes no separate note for Word for Mac, so the documented caveats are the ones listed above. Google Docs and browser converters run wherever the browser does, which is the usual answer when the Word on the Mac is an older version, or is not installed at all.
How can I convert a PDF file to a Word document?
Four routes cover almost everyone. Open it in Word itself, which converts on the way in. Upload it to Google Drive and open it with Google Docs. Export it from Acrobat or another paid desktop suite. Or use an online converter in the browser, which needs no install and no particular version of Word at the other end. Which one to pick depends on one thing: whether the PDF contains real text or is a scan.
How do I open a PDF file as a Word document?
In Word, choose File, then Open, then pick the PDF. Word tells you it will make a copy and convert its contents into a format it can display, and the original PDF is left unchanged. Microsoft notes that the converted document might not have a perfect page-to-page correspondence with the original, and that the conversion works best with files that are mostly text. Save the result as .docx when you are happy with it.
How to convert PDF to Word without paying free?
Free costs nothing extra for the ordinary case across all four routes above: Word's own importer, Google Docs and the free tier of most online converters all handle a short text-based file at no charge. The expense only shows up with scans, because recognizing characters off an image calls a model that is not free to run — every tool has to fund that somewhere, whether as a daily cap, a paid tier or your own key. Here the split is published: Basic mode is free and needs no account: 10 pages a day and 100 a month, rising to 100 pages a day and 1,000 a month once you sign in. For a text-based file that path never calls a model at all.
Why won't Word let me edit a PDF?
Two ordinary causes. The PDF may be open in a reader rather than in Word — a reader displays the fixed page and was never meant to reflow it. Or Word converted the file and the page came in as a picture, which happens when there is no text layer to work from. Microsoft notes that a PDF containing mostly charts or other graphics might show up as an image for the whole page.
How to convert non-editable PDF to editable?
A non-editable PDF is almost always a scan: the page is a photograph, so there are no characters in the file to edit. Character recognition has to read the image first. What matters next is where the recognised text lands. Dropped into floating boxes it falls apart the moment you type; rebuilt into real Word headings, paragraphs and table rows it behaves like a document you wrote yourself.
How do I convert a PDF to Word without messing up formatting?
Formatting survives only when it is rebuilt as Word constructs: a heading style, a real table object, a section carrying a column count. A page reproduced as positioned text boxes looks identical on open and stops behaving like a document on your first edit — press Enter inside a paragraph and watch whether the rest of the page moves. Whichever route you take from the four above, run that check on one page before you commit the whole file.