Etoolio

Image to Text (OCR)

Free image to text converter (OCR). Copy text from photos, screenshots and scanned PDFs in 28 languages including Arabic, Urdu and Hindi. Batch files, paste with Ctrl+V. Never uploaded.

Updated

Drop images or a scanned PDF here

Or paste a screenshot with Ctrl+V (Cmd+V on a Mac). JPG, PNG, WebP, BMP, GIF or PDF, up to 25 MB each. Nothing is uploaded.

Text recognition runs in your browser with the open-source Tesseract engine. Your images and PDFs are never uploaded. The first time you use a language, its data file (a few megabytes) downloads from the jsDelivr CDN and is then kept by your browser. Handwriting is not supported well; printed text, screenshots and scans work best.

Copy the text out of any picture

Text inside a photo, a screenshot or a scanned page is just pixels: you cannot select it, search it or paste it anywhere. This tool reads it with OCR (optical character recognition) and gives you plain text to copy or download. It runs the open-source Tesseract engine inside your browser, so your images are never uploaded, there is no daily limit and nothing to sign up for.

How to use it

  1. Pick the language of the text. If a document mixes two, such as English and Arabic, choose both.
  2. Add your images: drop them on the box, choose files, or copy a screenshot and press Ctrl+V (Cmd+V on a Mac). Up to 20 files at once, and scanned PDFs up to 30 pages.
  3. Copy or download the text. Tick "Join lines into paragraphs" to undo the line breaks from the page, which is handy when you paste into a document or email.

The first time you use a language, its data file (a few megabytes) downloads from the jsDelivr CDN; after that the browser keeps it and recognition starts straight away.

What it reads well, and what it does not

Works wellStruggles
Screenshots, documents, letters, book pagesJoined-up handwriting
Receipts, invoices, forms, ID cardsVery small or blurred text
Signs, menus and posters photographed straight onText on busy photos, or at a steep angle
Scanned PDFs at a normal resolutionDecorative fonts and text in tables with many lines
Arabic words in a clear fontArabic-Indic digits (٠ to ٩), which are often misread: check numbers

Getting better results

  • Choose the correct language: reading French with the English model drops the accents, and Arabic needs the Arabic model.
  • Photograph the page straight on, in even light, and crop away the background.
  • Use the original screenshot rather than a photo of a screen.
  • Watch the confidence score on each result. Above 85% the text is usually clean; below 60% check it line by line.

28 languages, right to left included

It reads English, Arabic, French, Spanish, German, Portuguese, Italian, Dutch, Russian, Ukrainian, Polish, Turkish, Greek, Hindi, Urdu, Bengali, Persian, Hebrew, Chinese (Simplified and Traditional), Japanese, Korean, Indonesian, Malay, Vietnamese, Thai, Filipino and Swahili. Arabic, Urdu, Persian and Hebrew results are shown right to left. To turn the text into another language afterwards, paste it into the translator.

Related tools

  • A PDF with selectable text is faster and exact with PDF to text, which also offers OCR for scanned pages.
  • A scanned bank statement can go straight into the bank statement converter, which now reads scans with the same engine.
  • iPhone HEIC photos open here in Safari; in other browsers convert them first with the image converter.

Good to know

Image to Text (OCR), frequently asked questions

How do I copy text from an image?

Drop the image here, choose a file, or paste a screenshot with Ctrl+V (Cmd+V on a Mac). Pick the language of the text and the tool reads it with OCR (optical character recognition). The text appears in a box you can copy or download as a .txt file.

Is this OCR really free?

Yes. There is no daily limit, no sign-up and no watermark, because the recognition runs on your own device with the open-source Tesseract engine rather than on a paid server.

Are my images uploaded?

No. Images and PDFs are read in your browser and never sent anywhere. The only download is the language data the first time you use a language (a few megabytes from the jsDelivr CDN), which your browser then keeps.

Can it read Arabic, Urdu or Hindi?

Yes. It reads 28 languages, including Arabic, Urdu, Persian, Hebrew, Hindi, Bengali, Chinese, Japanese and Korean, and you can pick two at once for a document that mixes, say, English and Arabic. Right-to-left text shows in the right direction. Arabic-Indic digits (٠ to ٩) are its weak spot, so check numbers in Arabic documents.

Can it convert a scanned PDF to text?

Yes. Drop the PDF and each page (up to 30) is rendered and read with OCR. If the PDF already has selectable text, the PDF to text tool is faster and exact.

Does it read handwriting?

Not well. Tesseract is trained on printed text, so screenshots, documents, receipts, signs and book pages work best. Neat block capitals sometimes work; joined-up handwriting usually does not.

How can I get more accurate text?

Choose the right language, use a sharp photo taken straight on with even light, and crop away anything that is not text. The confidence score on each result tells you how sure the engine is; below 60% it is worth checking the text carefully.