OCR a scanned PDF (English)

Scanned pages are just pictures. This tool renders each page, recognises the words with a WebAssembly build of Tesseract that ships with the site (no downloads from third parties, nothing uploaded) and writes the result as an invisible text layer over the original pages, or as plain text.

Loading tool…

How to use OCR PDF

Select a scanned PDF.
Choose the resolution (higher is slower but more accurate) and the output type.
Run OCR – recognition is CPU-intensive and can be cancelled.
Download the searchable PDF or text.

Supported formats

Input: PDF (scanned)
Output: Searchable PDF, TXT

How the searchable PDF is built

Your original pages are kept as they are. Recognised words are added as invisible text positioned over each word so you can search, select and copy.

Accuracy tips

Use 200–300 dpi for small print. Straight, high-contrast scans work best; handwriting, heavy skew and low-resolution images reduce accuracy.

Limitations

English only.
Accuracy depends on scan quality; handwriting is not supported.
Invisible text uses a basic Latin font, so accented characters outside Latin-1 become “?” in the text layer.
Large initial download for the OCR engine (about 15 MB) on first use; it is then cached by the browser.

Privacy

Files you choose are read into your browser’s memory and processed on your device; they are not uploaded to a server. Temporary download links are released after use, and the offline cache never stores your documents. More about privacy and limits.

Frequently asked questions

Which languages are supported?

English only. Other languages need additional language data that is not bundled.

Is it slow?

It runs on your CPU, typically several seconds per page. Close other tabs for large documents.

Is the file uploaded for recognition?

No. The OCR engine and English model are served from this site and run in your browser.