OCR PDF – recognize text in scans
A scan is only a picture of the page. Text recognition lays the recognized text invisibly over it: the document looks as before, and you can search it, copy text, change it or translate it.
Drop your file here or choose one
A scanned PDF or a photo (JPG, PNG)
Our Terms apply.
How it works
- Drop a scanned PDF or a photo into the box – it opens in your browser. Pages that already have text stay as they are.
- Choose the document’s language and press “Recognize text”. “Standard” reads in the browser; “Precise” has the AI read every page.
- Download, or go on to “Edit PDF” or “Translate PDF” without uploading again.
Frequently asked questions
Doesn’t a scanned PDF already contain text?
Usually not. A scanner or a phone app stores every page as a photo inside the PDF. You see letters, but to the computer it is an image: searching, copying, editing and translating only work after text recognition. Some scan apps recognize the text themselves; we leave such pages as they are.
Does the PDF look different afterwards?
No. The image of the page stays as it is; only a page scanned slightly askew (up to 5 degrees) is set straight. The recognized text lies invisibly right over the words. Only when you change a line in “Edit PDF” is that line covered in the paper’s color and written anew; everything else stays the scan.
Does my file leave my computer?
Not with “Standard”: recognition runs in your browser and nothing is uploaded. With “Precise” the images of the pages go to the AI of a provider with servers in the EU – not used for training, not stored. The PDF itself stays in your browser.
What is the difference between “Standard” and “Precise”?
“Standard” reads clean scans well, free and without an account, up to 20 pages (10 on a phone). “Precise” has an AI read every page again: better with poor scans, phone photos, small print and tables, up to 100 pages, with an account. The browser still finds where the lines are; the AI reads the words from the page image. “Precise” is not switched on everywhere yet; where it is missing, the tool says so. The result shows uncertain words.
What does it cost?
“Standard” is free, without limit, and runs in your browser. “Precise” needs an account and is not available at launch; accounts are coming later.
Which languages are recognized?
The 20 languages of “Translate PDF”: German, English, French, Spanish, Italian, Dutch, Portuguese, Polish, Czech, Swedish, Danish, Norwegian, Finnish, Romanian, Hungarian, Croatian, Greek, Russian, Ukrainian and Turkish. For mixed documents choose a second language.
In which font is a recognized line changed?
In one that matches the scan: text recognition measures whether a line is typewritten, bold or italic; with “Precise” the AI also spots serif type. When you change the line in “Edit PDF”, it is written in Courier or Times, for example, bold or italic as before. You can also choose font and style there yourself.
Does handwriting work?
Printed text is recognized reliably. Handwriting, stamps and signatures stay part of the image; “Precise” often reads clear hand printing too, but not reliably.
Related tools
Why Docuvoza
No account, no subscription, no document storage. Most tools run on your device; the rest in the EU, deleted after 60 minutes. Your working copy stays in this browser for up to 7 days, or until you choose Start over. And when someone has to sign: soon with Signovoza.
How the handover works →From a scan to an editable document
The lease from the folder, the invoice photographed with your phone, the form from the copier: as scans they are images. “OCR PDF” turns them into a PDF you can search and copy from – and change in “Edit PDF” or bring into another language in “Translate PDF” without uploading again.
Like Adobe, without a subscription
Open a scan in “Edit PDF” or “Translate PDF” and text recognition is offered right there. Then you change a line as in any PDF: only that line is written anew, in a matching font and color; the rest stays the scan.
What the result shows
How many pages and words were recognized, which pages had text already, and where words are uncertain. A click on such a page opens it for checking. Before you start, + and − enlarge the preview (50 to 200%).
Your file stays with you
“Standard” runs entirely in your browser. With “Precise” only the images of the pages go to the AI on EU servers, page by page, and are not stored.