How to convert a scanned PDF to Word with OCR

Confirm which pages need recognition

Try selecting a sentence or searching for a distinctive word. If neither works, the page may be an image. Some documents mix scans with digital pages; OCR PDF can skip pages that already contain text. Keep the original scan to check the recognized wording later.

Prepare the scan and select languages

Use straight, clearly lit pages without shadows or clipped edges. In OCR PDF, choose the language present on the page, not simply the interface language. Select both languages for a bilingual document, but avoid adding unnecessary languages. The 300 DPI option helps small print at the cost of extra time and memory; 200 DPI can suit larger type.

Latin-script scans: searchable PDF, then Word

For supported Latin-script languages such as English, French and Spanish, choose searchable PDF as the OCR output. Run recognition, download that PDF, then open PDF to Word and convert this new file to .docx.

• Check that a word can be selected in the OCR result.
• Convert the searchable PDF, not the original image-only scan.
• Open the downloaded Word document and correct the text and paragraph breaks.

Arabic and Russian: use the text download

This OCR tool recognizes Arabic and Russian but does not create a searchable PDF text layer for these scripts. Download the recognized .txt instead, open it as UTF-8 text, and paste it into Word or another word processor. Set the correct text direction, rebuild headings and save as .docx.

Do not pass the unchanged image-only PDF to PDF to Word and expect the recognized words to appear. The text-file route is a manual editing workflow, not an automatic layout-preserving Arabic PDF-to-DOCX conversion.

Proofread before using the document

Compare names, dates, amounts and negations with the scan. OCR may confuse 0 with O or omit faint punctuation. Columns and tables may be read in the wrong order, so check them separately. PDF to Word rebuilds text flow; it does not preserve exact fonts, pictures, tables or complex page layout.

OCR and PDF to Word run in your browser without uploading the document. OCR downloads its engine and languages from a public CDN, which can see ordinary request information but not your pages. Long scans can be slow on phones: process a short sample first rather than assuming a large document will fit in memory.