A scanned PDF looks like a document but behaves like a photograph: you cannot select its words, search them, or fix a typo, because there are no words in the file — only a picture of words. Converting it to an editable Word document therefore takes an extra step called OCR (optical character recognition), software that reads the picture and rebuilds real text. The good news is that capable OCR is available free, including from Google and Microsoft.
This guide covers the three free routes worth using — Google Docs (best all-rounder), Microsoft Word itself (best when you have it), and dedicated free OCR tools for difficult files — plus the accuracy habits that separate a clean conversion from a garbled one. A quick test before you start: open the PDF and try to select a sentence with your mouse. If you can select and copy text, your PDF is already digital — skip OCR entirely and just open it in Word (File > Open) or convert it directly.
- Google Docs (free, no software): upload the PDF to Google Drive, right-click > Open with > Google Docs, then File > Download > Microsoft Word (.docx).
- Microsoft Word (if you have it): File > Open the PDF and let Word convert it.
- Free OCR websites: for files Google Docs rejects or mangles — with the privacy caution below.
- Whatever the route: proofread names, numbers and dates against the original. OCR is good, not perfect.
What OCR does, and its honest limits
OCR software analyzes the shapes in a scanned page, matches them to characters, and outputs a document with real, editable text. Google's own documentation for its Drive conversion states the practical limits clearly, and they apply to every OCR tool:
- File size: Google Drive converts PDFs and photos to text only when the file is 2 MB or smaller — bigger scans must be split or compressed first (our PDF compression guide covers that).
- Quality in, quality out: Google advises using files that are the right way up, with sharp focus and clear contrast — a shadowed, skewed phone photo converts poorly everywhere.
- Formatting is approximate: Google states plainly that bold, italics, font sizes and line breaks may not be retained, and that lists, tables and columns are unlikely to survive. Expect to rebuild tables by hand.
- Handwriting is the hardest case: neat printed-style handwriting sometimes converts; cursive usually does not, in any free tool.
Set expectations accordingly: OCR gives you a fast, editable first draft of the text — a genuine time-saver over retyping — not a perfect replica of the page.
Method 1: Google Docs (free, works anywhere)
- Go to drive.google.com and sign in with any Google Account.
- Click New > File upload and choose your scanned PDF. (If it is over 2 MB, split it into smaller PDFs or compress it first — Drive will upload a big file happily but refuse to convert it.)
- Once uploaded, right-click the file > Open with > Google Docs. Google runs OCR and opens an editable document: the scanned page image at the top, the recognized text below it.
- Delete the embedded page image if you do not need it, tidy obvious errors, and re-apply any structure (headings, lists) that did not survive.
- Choose File > Download > Microsoft Word (.docx) to get your Word document. (Keeping it in Docs is fine too if Word format is not required.)
This route handles printed English and most major languages well, costs nothing, and needs no installation — for a straightforward scanned letter, form or book chapter it is the one to try first. Note that you are uploading the document to Google; for highly sensitive files, prefer the Word method below, which works locally.
Method 2: Microsoft Word itself
If you have the desktop Microsoft Word app (Microsoft 365 or a recent standalone version), it converts PDFs directly:
- Open Word and choose File > Open, then browse to the PDF.
- Word shows a message that it will convert the PDF into an editable Word document — click OK.
- Review the result, then Save As a .docx.
Microsoft's own caveat matters here: Word's conversion “works best with PDFs that are mostly text” — digitally created documents. For a true scan, Word attempts recognition too, but results on photographed or skewed pages are usually weaker than Google's OCR, and heavily formatted pages (columns, tables) can come out rearranged. Treat Word as the first try for text-based PDFs and for sensitive documents you do not want to upload anywhere, and fall back to Google Docs or a dedicated OCR tool when the output disappoints.
Word for the web (the free online version) does not convert uploaded PDFs the same way, so this method needs the desktop app. No Word and a sensitive document? A desktop OCR app trial, or your scanner's bundled software, keeps everything on your own machine.
Method 3: Free OCR tools and built-in helpers
When Google Docs refuses a file (too big, too many pages) or mangles it, dedicated OCR services are the fallback. Well-known free options such as OnlineOCR and iLovePDF's OCR work the same way: upload the PDF, choose the language, run OCR, download a .docx. Free tiers limit pages or files per day — enough for occasional documents.
Privacy caution: like online compressors, OCR websites process your document on their servers. Do not upload IDs, financial records, medical or legal documents to them. For those, use the desktop Word method or another on-device tool.
Two built-in helpers worth knowing for short excerpts:
- Windows Snipping Tool has a “Text actions” feature that copies text from a screenshot — open the PDF, snip the paragraph, copy the text, paste into Word. Ideal for a page or two.
- Microsoft OneNote (free) can “Copy Text from Picture”: insert the scanned page as an image, right-click, and paste the recognized text into Word.
- On a Mac, Preview and Photos offer Live Text on recent macOS versions — select the text right on the scanned page and copy it across.
Accuracy tips: get a clean conversion the first time
- Scan or photograph well: flat, straight, evenly lit, no shadows or fingers. 300 dpi is the OCR sweet spot; a blurry 600 dpi photo converts worse than a sharp 300 dpi scan.
- Straighten and crop before converting — most phone scanner apps do this automatically in “document” mode.
- One language per document, and pick that language in the tool when asked; mixed-language pages confuse recognition.
- Split big files: convert chapter by chapter (staying under Google's 2 MB limit) rather than fighting one huge upload.
- Proofread the high-risk items: names, dates, totals, account and ID numbers, decimals. OCR's classic errors are look-alike characters — 0/O, 1/l, 5/S — exactly where accuracy matters most.
- Rebuild tables manually: since tables rarely survive conversion, paste the recognized text into a fresh Word table rather than untangling a broken one.
Working with several related PDFs — say, converting chapters and then merging the originals for reference? Our guide to combining PDFs on a Mac without Adobe covers the merge side.
Key takeaways
- A scanned PDF is a picture of text; OCR is the recognition step that makes it editable — test by trying to select text in the PDF.
- Google Docs is the best free route: upload to Drive, Open with Google Docs, download as .docx — but files must be 2 MB or under, and tables/columns rarely survive.
- Desktop Word converts PDFs itself and keeps everything on your machine — best for text-based PDFs and sensitive documents.
- Free OCR websites handle stubborn files, but you are uploading the document — never use them for IDs, financial or legal papers.
- Sharp, straight, 300 dpi source pages and a proofread of names and numbers are what separate a clean conversion from a mess.
Quick questions
Why did Google Docs give me a document with the scanned image and text below it?
That is normal Google behavior: it keeps the original page image at the top of the converted document and places the recognized text underneath. Delete the image once you have checked the text against it, and you are left with a clean editable document.
Can OCR handle handwriting?
Only sometimes, and free tools struggle most here. Neat, printed-style handwriting on a clean page may convert passably; joined-up cursive usually produces errors on most lines. For handwritten material, OCR output is a starting draft at best — expect heavy correction, or transcribe short documents manually.
The conversion lost my tables and layout. Is there a better tool?
Every OCR tool — free or paid — treats layout as secondary; Google states openly that lists, tables and columns are unlikely to be retained. Paid desktop OCR (such as ABBYY or Acrobat Pro) preserves layout better, but for a one-off document it is faster to rebuild the table in Word from the converted text.
Is it legal to convert a scanned document I did not create?
Converting documents you own or have permission to use — your own contracts, letters, forms — is unproblematic. Copyrighted books and articles are different: converting a personal copy for accessibility or study is treated differently from distributing the text. When in doubt, convert for your own use and do not share the result.
Further reading: Convert PDF and photo files to text (OCR limits and quality notes) — Google Drive Help.