If you can click a scanned page in Microsoft Word but cannot place the cursor between its words, Word contains an image, not editable text. OCR can extract characters from that image, but you still need to create a text document and proofread it before treating it as a usable Word file.
Separate text extraction from rebuilding a document
OCR answers a narrow question: “Which characters are visible in this image?” It does not promise that a heading will remain a heading, that a table will become a Word table, or that a stamp, signature, footnote, and two-column layout will land in the right place. Treat the original scan as the record and the OCR output as a draft for review.
Microsoft documents OCR in OneNote for Microsoft 365, OneNote 2024, OneNote 2021, and OneNote 2016. You can copy recognized text from a picture or a file printout and paste it into Word. Microsoft also warns that availability depends on the image’s complexity, legibility, and amount of text; recognition may not be immediate, and some results can take 24–48 hours. That is a OneNote workflow, not a claim that Word itself converts every scanned page into an editable, layout-faithful document.
Make a reviewable Word draft
Start with a copy of the scan, not the only original. If the page contains confidential or regulated information, use an approved process and do not upload it merely for convenience. Then use one of these two routes:
- In a supported desktop OneNote version, insert the picture or file printout, use its Copy Text command when available, and paste the result into a new Microsoft Word document.
- For an image you are authorized to process, use ToolboxHub OCR (Image to Text) to recognize the image in the browser, copy the plain text, and paste it into a new Word document.
ToolboxHub OCR accepts image files, lets you select a recognition language, and returns plain text that you can copy or download as a .txt file. It does not open, edit, or save DOCX files; it does not recreate Word styles, page breaks, tables, columns, or the visual layout of the scan. If your source is a multipage PDF, convert only a permitted page to an image first or use an approved document workflow. Image to PDF creates a PDF from images; it is not an OCR or Word-conversion feature.
For Word on the web, Word for Mac, or a desktop build whose menus differ, use the same boundary: obtain recognized text through a supported OCR route, paste into a new document, and compare against the scan. Do not assume the OneNote right-click menu exists in every app or platform because it exists in the documented desktop OneNote versions.
Proofread the failure-prone parts first
Read the scanned page beside the Word draft. Start with characters that can change a decision: names, dates, amounts, account references, addresses, negations, and headings. Then inspect layout-sensitive material such as tables, checkboxes, marginal notes, superscripts, and multi-column text. A line that looks complete in a plain-text OCR result may have been assembled from the wrong column.
For a table, rebuild the table in Word manually after you have checked each cell. For a form, preserve the scan and mark the OCR copy as a working transcription. For a document that needs legal, financial, medical, or contractual accuracy, have an appropriate reviewer compare the final Word document to the source; OCR is not evidence that every word was captured correctly.
Before sending the new document, run a practical post-use check: search for a distinctive name, compare one number in every important row, make sure page headers were not copied into body text, and read the Word draft in Print Layout. If you have a word-count requirement after correction, Word Count can count pasted text, but it cannot validate the scan or a DOCX file.
Common questions
Why can’t I edit text in a scanned Word page?
The scan is commonly embedded as an image. It looks like a page of writing, but Word has no individual text characters to edit until an OCR process extracts them.
Does OCR preserve Word formatting and tables?
No. OCR may identify text, but it does not reliably reconstruct styles, columns, table cells, handwriting, stamps, or page geometry. Rebuild and verify those elements in Word.
Can I use the OneNote instructions in every Word version?
No. Microsoft’s cited OCR article applies to named OneNote desktop versions, not all Word, web, or Mac interfaces. Use the supported route available to you and verify the pasted result against the scan.
Does ToolboxHub edit my DOCX after OCR?
No. ToolboxHub OCR processes a selected image and produces plain text in the browser. It neither opens nor edits DOCX files, and it cannot preserve a scanned Word page’s layout.