How to convert an image to text
- Add your images: drop them in the box, click to choose files (on a phone you can take a photo), or paste a screenshot with Ctrl+V.
- Choose the language of the text and click Extract text.
- Check the result in the text box and fix any mistakes.
- Download Word (.docx), download a .txt file, or copy the text.
Several images are processed in order, which is perfect for a document photographed page by page.
What is OCR?
OCR (optical character recognition) is technology that recognizes letters and words in a picture and turns them into real, editable text. This tool uses Tesseract, a widely used open-source OCR engine, running directly in your browser. Your images stay on your device.
Tips for the most accurate results
- Photograph straight on: hold your phone parallel to the page to avoid slanted lines.
- Use good, even light: avoid shadows and glare across the text.
- Get close: the text should fill the frame. Larger text is recognized far more accurately.
- Crop out the background and any pictures, so only the text area remains.
- Pick the right language so accents and special characters are recognized.
What you can use it for
| Use | Example |
|---|---|
| Retype-free documents | Turn a printed letter or form into an editable Word file |
| Study notes | Copy text from textbook pages or lecture slides |
| Screenshots | Grab text from an image, a video frame or an app that won't let you copy |
| Receipts and invoices | Pull out amounts and details for expense reports |
| Quotes and articles | Copy text from a photo of a book, sign or poster |
Limitations
OCR is very good with clear printed text but isn't perfect. Results are usually 95%+ accurate on clean scans, lower on photos of curved pages, low-light images, handwriting or decorative fonts. It extracts text only; tables, columns and images aren't rebuilt. Always proofread important documents.
Frequently asked questions
Are my images uploaded to a server?
No. Text recognition runs inside your browser on your own device. Your images are never sent to us. The first time you use a language, your browser downloads that language's recognition data (a few MB), which is then cached for next time.
What kinds of images work best?
Clear, well-lit photos or screenshots of printed text work best: straight on, in focus, with good contrast and text at least 20 pixels tall. Scans at 300 DPI give excellent results. Blurry, angled or very small text, and fancy fonts, reduce accuracy.
Can it read handwriting?
Only neat, printed-style handwriting, and with lower accuracy. This OCR engine is designed for printed and typed text. Always check and correct the result before using it.
How do I convert an image to a Word document?
Add your image, click Extract text, review the text in the box, then click Download Word (.docx). The file opens in Microsoft Word, Google Docs, LibreOffice and Pages.
Does it keep the original formatting?
It keeps the line and paragraph breaks, but not fonts, columns, tables or images. For documents with complex layouts, you'll need to tidy the formatting in Word afterwards.
Can I convert several images at once?
Yes. Add multiple images and they're processed in order, with the text from each one added below the last — handy for multi-page documents photographed page by page.