What Is OCR?
OCR (optical character recognition) is technology that recognises letters and numbers in an image and turns them into real, editable text. A photo of a document, a screenshot, a scanned page or a picture of a sign contains only pixels: you can't search it, copy from it or edit it. OCR reads those pixels and gives you text you can paste into an email, a Word document, a spreadsheet or a translation tool. This image to text converter uses Tesseract, a highly accurate open-source OCR engine originally developed by HP and later maintained with support from Google, running directly in your browser.
How to Extract Text From an Image
- Add your images. Drag and drop them, click to choose files, or paste a screenshot with Ctrl+V (Cmd+V on Mac). Scanned PDFs work too.
- Choose the language of the text. Selecting the right language makes a big difference to accuracy, especially for accented letters.
- Click Extract text. The first time you use a language, its recognition model (a few megabytes) is downloaded and then cached by your browser.
- Check and correct the text in the result box, then copy it or download it as a .txt file or a searchable PDF.
Supported Languages
You can recognise text in English, German, French, Spanish, Italian, Portuguese, Dutch, Polish, Turkish, Swedish, Danish, Norwegian, Finnish, Czech, Romanian, Hungarian, Greek, Russian, Ukrainian, Arabic, Hindi, Bengali, Urdu, Chinese (Simplified and Traditional), Japanese, Korean, Thai, Vietnamese and Indonesian. Select up to three languages at once for documents that mix languages, such as an English and Hindi form, or a German document with English terms.
Tips for the Most Accurate Results
| Tip | Why it helps |
|---|---|
| Use good, even lighting and avoid shadows | OCR works by contrast between text and background |
| Hold the camera straight above the page | Tilted or curved lines are harder to read |
| Crop to the text you need | Less background means fewer stray characters |
| Keep "Enhance image" on for photos | Greyscale, upscaling and contrast boost help small or faint text |
| Choose the correct language | The engine uses a dictionary and character set for each language |
| Use high-resolution scans (300 DPI) | Tiny text in low-resolution images loses detail |
Printed text in clear fonts is usually recognised with more than 95% accuracy. Handwriting, decorative fonts, very small text, low-contrast images and complex layouts (tables, multiple columns) are harder, so always proofread the result. The confidence score gives you a quick sense of how reliable the recognition was.
What Is a Searchable PDF?
A searchable PDF looks exactly like your original image or scan, but has an invisible layer of recognised text behind it. That means you can search it with Ctrl+F, select and copy text, and have it indexed by document management systems, while keeping the original appearance. Tick "Also create a searchable PDF" before extracting, and a Download searchable PDF button appears with all your pages combined into one file.
Common Uses
- Copy text from screenshots of apps, videos, presentations or error messages.
- Digitise paper documents: letters, receipts, invoices, recipes and notes.
- Quote from books and articles without retyping.
- Make scanned PDFs searchable for archives and research.
- Translate text in photos: extract it here, then paste it into a translator.
- Edit scanned forms: extract the text, or fill the form with our Fill PDF Form tool.
Private OCR in Your Browser
Many online OCR services upload your images to their servers, which isn't ideal for ID cards, contracts, bank statements or medical letters. Here, recognition happens in your browser using WebAssembly. The only things downloaded are the OCR engine and the language models, never your images. After the first use, the models are cached, so recognition starts faster and even works offline.
Frequently Asked Questions
Is this OCR tool free?
Yes. There's no sign-up, no watermark and no limit. Process up to 20 files at a time and as many batches as you like.
Can it read handwriting?
Neat, printed-style handwriting can sometimes be recognised, but accuracy is much lower than for typed text. Joined-up handwriting usually isn't readable.
Can I extract text from a PDF?
Yes. Scanned PDFs are processed page by page (up to 30 pages). If your PDF already contains selectable text, the PDF to Text tool is faster and exact.
Why is the first run slower?
The OCR engine and language model are downloaded the first time. After that they're cached by your browser.
Are my images uploaded?
No. Text recognition runs entirely on your device.