Steps
Recognise the text and save a searchable PDF
Drop the scanned PDF and choose the document's language, plus a second language for mixed documents. Also create a searchable PDF is ticked for you. Leave PDF pages blank to process every page; only the pages you list go into the searchable PDF. Press Recognise text, check the recognised text and the average confidence, then press Download searchable PDF. The first run downloads the OCR engine and language data, which your browser caches. Long scans take a while, especially on phones; keep the tab open.
Convert the searchable PDF to Word
Drop the searchable PDF from the previous step (not the original scan) and keep Editable text selected. Leave Join lines into paragraphs and Detect headings on so sentences flow and larger lines become headings. Press Convert to Word, check the preview and download the .docx. Tables and photos on the scan are rebuilt as Word tables and pictures; check them against the scan. Proofread names, dates, amounts and reference numbers against the scan before you use the document.
Requirements and sources
| What | Value | Source |
|---|---|---|
| Microsoft Word - which PDFs convert best | Converting from PDF to Word works best with files that are mostly text—for example, business, legal, or scientific documents. | Opening PDFs in Word - Microsoft Support Last verified 2026-10-21 |
| Microsoft Word - converted layout | But when you open a PDF file in Word, it might not look exactly the way it looked as a PDF. | Opening PDFs in Word - Microsoft Support Last verified 2026-10-21 |
| Microsoft Word - pages that are mostly graphics | If the PDF contains mostly charts or other graphics, the whole page might show up as an image. | Opening PDFs in Word - Microsoft Support Last verified 2026-10-21 |
| Google Drive OCR - file size | The file should be 2 MB or smaller. | Convert PDF and photo files to text - Google Drive Help Last verified 2026-10-21 |
| Google Drive OCR - text size | Text should be at least 10 pixels high. | Convert PDF and photo files to text - Google Drive Help Last verified 2026-10-21 |
| Tesseract OCR engine - languages | Tesseract has unicode (UTF-8) support, and can recognize more than 100 languages "out of the box". | tesseract-ocr/tesseract - GitHub (project README) Last verified 2026-10-21 |
FAQ
Why does my scanned PDF convert to an empty Word file?
Because the pages are images: there is no text inside the PDF for a converter to read. The PDF to Word step says so when it finds no text. Run OCR first, then convert the searchable PDF it produces.
Which languages can the OCR read?
Our tool uses Tesseract and offers about 30 languages in its list, including Chinese, Japanese, Arabic and Hindi, and you can add a second language for mixed documents. Tesseract itself supports more than 100 languages. Each language's data is downloaded the first time you use it.
Can it read handwriting?
Not reliably. Tesseract is designed for printed text. Neat block capitals sometimes work; cursive handwriting does not.
How do I get fewer OCR mistakes?
Start from the sharpest scan you have: straight pages, no shadows, and text large enough (Google's Drive help recommends text at least 10 pixels high for its own OCR, a useful rule of thumb). Choose the right language, and check common confusions such as 0 and O or 1 and l. Rescanning at a higher resolution usually helps more than any setting.
Will the Word file look like the original page?
Not exactly. In Editable text mode you get paragraphs and headings that are easy to restyle, but multi-column layouts, tables and forms lose their structure. If you need the exact look rather than editable text, PDF to Word's Exact look option places each page as an image, which cannot be edited.
BufferMint is not affiliated with Microsoft, Google or the Tesseract project; their pages are cited for comparison and general guidance. Files are processed in your browser and are not uploaded.