OCR PDF
Add an invisible text layer to scanned pages so you can search, select and copy.
next step PDF to textCompress PDF
report a problemHow it works
- 01Pick the document's language and press "Download & start" once, then drop in a scanned PDF.
- 02Pages that already have text are left alone; the rest are read one by one, with progress per page.
- 03Download the searchable PDF, which looks exactly like the original, or just its text as a .txt file.
Questions
›Will my PDF look different?
No. The original pages are kept as they are, scan quality included. The recognized text is laid invisibly on top of each page, word by word in the right place, so you can search it, select it and copy it, and screen readers can read it. The file grows a little (usually a few KB per page).
›Which OCR model is used, and what is its license?
Tesseract, the open-source OCR engine developed by Google and its community, with its "fast" neural-network language models (tessdata_fast), both under the Apache-2.0 license. It writes the invisible text layer itself, in a way that works for every script, including Chinese, Arabic and Hindi. Each language is a one-time download of 0.5–6 MB plus about 4 MB for the engine.
›How long does it take?
Usually 3–8 seconds per page on a laptop, longer on phones and for dense pages. Each page is rendered at 300 dpi and read on your processor; on computers with many cores two pages are read at once. You can stop at any time and still download the pages read so far.
›How accurate is the text?
Very good on clean, straight scans of printed text. Low-resolution scans, faint copies, handwriting and photos taken at an angle give more mistakes. It's AI, so check anything important; the text you can copy is exactly what was recognized. Picking the right language matters most.
›Are my files uploaded anywhere?
No. Your files are processed on your own device by your browser. They never touch our servers, so there are no upload waits, no size caps from us, and nothing to delete afterwards.