Extract text from any PDF — including scanned documents — using OCR. Downloads as an editable Word document (.rtf).
Works on text PDFs and scanned image PDFs. Output is .rtf
How it works:
Text-based pages have their text extracted instantly via PDF.js.
Scanned or image pages are rasterized then read with OCR using Tesseract.js (English).
The tool automatically detects which method to use for each page.
OCR pages take longer — please be patient for large scanned files.
Each page of your PDF is checked individually and the right extraction method is chosen automatically. Pages with real, embedded text are read directly using PDF.js — instant and highly accurate, since the text is already stored as text. Pages that are actually scanned images are run through Tesseract.js, an in-browser OCR engine that recognizes text visually. A single document can mix both page types; you don't need to know in advance which is which. The result downloads as a .rtf file that opens in Word, Google Docs, and LibreOffice.
Text pages and scanned pages are identified and handled correctly without any manual input from you.
Tesseract.js reads scanned or photographed text directly in your browser — no server-side processing.
Output is a standard .rtf file, compatible with Word, Google Docs, and LibreOffice.
Pages with real embedded text convert essentially perfectly. OCR on scanned pages is generally good but depends on scan quality — clean, high-resolution scans convert more reliably than blurry photos or faxes.
The .rtf format prioritizes editable text over pixel-exact layout, so complex multi-column or table-heavy pages are simplified rather than perfectly recreated.
No — this tool focuses specifically on extracting text content; embedded images are not carried into the output.
OCR effectively "reads" each scanned page visually, which takes more processing than copying existing text. There's a 300-page limit for this reason — larger scanned documents should be split first.