OCR PDF

What is OCR?

OCR (optical character recognition) looks at the picture of each page in a scanned PDF and works out which letters and words are in it. HerePDF then adds those words to the PDF as an invisible text layer, so you can search the file, select text and copy it. The pages themselves are not changed: they look exactly as they did before.

Is the result accurate?

Not always, and HerePDF does not claim any accuracy figure. OCR reads pixels, so a blurry scan, a skewed page, small print or a busy background can produce wrong or missing words. Each page is labelled good, check this page, too poor to add text or no text found, based on the engine's own estimate of how sure it is. That estimate is a guide, not a check that the words match the page, so read anything important yourself.

Does a text layer verify what the scan says?

No. The text layer is the engine's best reading of the picture, nothing more. A searchable PDF is not a certified or checked copy of the document, and the words in it can differ from the words on the page. The picture is still what a reader sees.

Can it read handwriting?

No. It is built for printed text. Handwriting usually comes out as wrong words or nothing at all, and pages like that are labelled too poor to add text or no text found.

Which languages are supported?

English, French, German, Spanish, Italian, Portuguese and Dutch for now. These all use the Latin alphabet. Other scripts, such as Cyrillic, Greek, Arabic or Chinese, Japanese and Korean, are not available yet. Pick the language the document is written in before you start; reading a French page as English gives worse results.

Is my PDF uploaded?

No. Recognition runs in your browser, and your PDF never leaves your device. The first time you use the tool, your browser downloads the recognition engine (roughly 1.5 MB) and the language data for the language you chose (the size is shown next to each language), so a first run is about 2 to 5 MB depending on the language. These files come from HerePDF itself, never from a third party. Your PDF is never uploaded and never stored. The engine and language files (not your PDF) are kept in your browser's cache storage, by the browser's normal cache and by HerePDF's offline copy, so later runs usually do not download them again. Clearing the site's data removes them, and the browser can clear them at any time.

Why does a big file take longer?

Each page is drawn as a picture and read in your browser, using your own computer, so a long document or a large page takes longer. HerePDF does not quote a speed, because it depends on your device. Keep the tab open until it finishes. You can cancel at any time, and nothing is kept from a cancelled run.

What does "Skip pages that already have text" do?

Pages that already carry a text layer are left alone and not read again, which saves time and avoids doubling up text. Switch it off if the existing text layer is wrong and you want the page read again.

What is the difference between this and Extract Text?

Extract Text only copies text that is already inside the PDF. A plain scan has none, so it comes out empty. OCR PDF reads the picture of the page instead, and can add a text layer to the PDF or give you the words as a .txt file.