OCR PDF
OCR scanned PDFs to add an invisible text layer, making recognized text searchable, selectable, and copyable. Page images and sizes stay unchanged.
Your file is uploaded over a secure connection and processed on our server. The original is deleted right after processing, the result after 1 hour (24 hours in your account).
Upload PDF files (up to 10), the same settings apply to each.
Choose up to three document languages under “Document languages.” The default is the page language plus English. Select every language present for better recognition.
Optionally enable “Auto-rotate pages” and “Straighten tilted pages (deskew).” Under “Pages that already have text,” keep “Skip” to leave pages with real text unchanged, or choose “Redo OCR” to replace an old OCR layer. Redo OCR does not support straightening.
Enter the PDF password if prompted, then select “Start OCR.” Enable “Also save the text as a .txt file” for an additional text output.
Each searchable PDF is named with the original filename followed by -ocr.pdf. If selected, the separate text file uses the original filename with .txt. For multiple PDFs, a ZIP named ocr-pdf.zip is also offered.
Guests can process PDFs up to 50 MB and 20 pages, with two simultaneous jobs. Registered users can process PDFs up to 100 MB and 100 pages, with four simultaneous jobs. Recognition takes about 2–10 seconds per page. Accuracy depends on scan quality, and handwriting and complex tables are not recognized reliably.
No account is required, and results have no watermark. Files are uploaded over HTTPS, processed on the server, and the original is deleted after processing. Digitally signed PDFs can be recognized, but OCR makes the signature invalid. Password-protected PDFs require a password, and the resulting PDF is saved without it. Passwords are not stored.
FAQ
What happens if every page already has text?
The tool reports that nothing was recognized. If you want to replace an old OCR layer, select “Redo OCR” and run recognition again. This replaces an old OCR layer, not real text.
What information appears in the results?
The results report how many words were recognized and on how many pages, and show a preview of the first page's text. The preview is built in your browser from the downloaded result, and the recognized text is not stored separately on the server.
Can I retrieve results later?
Registered users can download results again from processing history during the 24-hour retention period. History stores file names, sizes, and settings, but not the documents, recognized text, or passwords.
Can I OCR a JPG or PNG directly?
No. The input format for this tool is PDF, so an image file must first be made into a PDF using another method.