Make a scanned PDF searchable (OCR)
Add a searchable text layer to scanned PDF pages with a selected OCR language.
To make a searchable PDF from a scan, upload the scanned file, choose the language of the text and download a copy with an invisible text layer added by OCR. You can then select, copy and search the words while the page still looks like the scan. The tool accepts PDFs up to 25 MiB and 50 pages, supports ten OCR languages and can straighten tilted pages. It runs in a queue, so large files may take a while.
How to make a scanned PDF searchable
Choose a PDF, select the text language, and leave Straighten scanned pages ticked unless the pages are already straight. Submit the form. Because OCR is slow, the job goes into a queue and a status page shows progress. When it finishes, download the file named with -searchable. The scanned look is kept; only a hidden text layer is added.
Languages and straightening
Pick the language the document is written in: English, Spanish, French, German, Portuguese, Simplified Chinese, Hindi, Arabic, Japanese or Russian. One language is used for the whole file, so mixed-language documents will be less accurate. Straightening corrects slightly crooked scans, and pages that are sideways are rotated automatically, which usually improves recognition.
Limits and accuracy
A file can have up to 50 pages, pages are checked against a pixel limit, the output cannot exceed 100 MB, and a job is cancelled if it runs too long. Pages that already have real text are skipped rather than redone. OCR is never perfect: handwriting, poor scans and small print produce errors, so proofread anything important.
What to do with the result
Search the file with your viewer's find command, or extract the words with PDF to Text. For an editable document, send the searchable file through PDF to Word. If the output is large, Compress PDF can shrink it, and Split PDF can break a long scan into chapters to stay under the 50-page limit.
Frequently asked questions
How do I make a scanned PDF searchable?
Upload the scan, pick the language of its text and submit. The tool runs OCR in the background and gives you a PDF that looks the same but has a hidden text layer you can search and copy from.
Is OCR 100% accurate?
No. Clean, straight, printed text is usually recognised well, but handwriting, low-resolution scans, stamps and small fonts cause mistakes. Proofread numbers, names and anything else where one wrong character would matter.
How many pages can I OCR at once?
Up to 50 pages per PDF, with a 25 MiB upload limit. For a longer document, use Split PDF to cut it into parts of 50 pages or fewer, run each part, and join them again with Merge PDF.
Which languages does searchable PDF support?
English, Spanish, French, German, Portuguese, Simplified Chinese, Hindi, Arabic, Japanese and Russian. Choose one language per file; text in any other language within the same document will not be recognised well.
Why is my job waiting instead of finishing right away?
OCR takes much longer than other PDF tools, so jobs are processed from a queue, and a status page shows where yours stands. If it takes too long it is cancelled and you can try a smaller file.
Related tools
PDF to Text
Extract plain text from any PDF file in seconds.
Use tool →PDF to Word
Turn PDFs into fully editable Word documents.
Use tool →Merge PDF
Combine multiple PDFs into a single document.
Use tool →Compress PDF
Shrink PDF file size while keeping quality.
Use tool →Split PDF
Extract page ranges out of a PDF into separate files.
Use tool →Rotate PDF
Rotate all or specific pages of a PDF 90/180/270 degrees.
Use tool →