How do I copy text from an image?
- 1Choose one or more pictures (JPG, PNG, WebP, HEIC) or a scanned PDF, or drop them on the box.
- 2Pick the language the text is written in.
- 3Press Get text. The first time, the reader downloads its language data; after that it starts at once.
- 4Correct anything in the text boxes, then copy the text or download it as a .txt or Word file.
How does image to text (OCR) work?
- 01Picture → upright (phone EXIF turn applied) → up to ~12 megapixels
- 02PDF page → drawn at 250 dpi → read like a picture
- 03Tesseract OCR: finds text lines → recognises each word with its language model → confidence 0–100 %
- 04Reading confidence = average of the word confidences
Optical character recognition (OCR) looks at the shapes in a picture, finds the lines of text and matches each word against a trained model of the chosen language. This tool uses Tesseract — the open-source OCR engine first developed at HP and later improved by Google — compiled to run inside your browser with Tesseract.js.
Accuracy depends most on the picture: straight, sharp, evenly lit text of normal size reads almost perfectly; blurry, curved or very small text and fancy handwriting read less well. Choosing the right language matters too, because the model also knows the language's letters and common words — which is why Urdu, Arabic and Hindi each need their own model.
Nothing is uploaded. Only the reader's program and the language data are downloaded from a public CDN the first time; your pictures are read on your own device, so you can safely use it for ID cards, receipts, contracts and notes.
Good to know
- Photograph the page straight on, in good light and filling the frame — straight, sharp text reads best.
- Pick the right language: an Urdu page read as English gives nonsense.
- Check numbers, names and amounts before you use them; the confidence figure shows how sure the reader was.
- For a whole scanned PDF that you also want to edit, use the PDF editor: it reads the pages and lets you change the text in place.
- Screenshots of websites and apps read almost perfectly — handy for copying text you can't select.
Image to text — common questions
Is image to text free and private?
Yes. It's free, needs no account, and your pictures are read in your browser on your own device. Nothing is uploaded; only the reader and its language data are downloaded the first time.
Which languages can it read?
English, Spanish, French, German, Portuguese, Indonesian, Italian, Dutch, Turkish, Russian, Hindi, Bengali, Arabic and Urdu. Choose the language the text is written in before you press Get text.
Can it read handwriting?
Neat, printed-style handwriting often reads well; joined-up or messy handwriting usually doesn't. The reader is trained mainly on printed text.
Can I convert a scanned PDF to text?
Yes. Choose the PDF and each page is read like a picture (up to 40 pages at a time). Download the result as a .txt or Word file.
How accurate is it?
For clear printed text, usually above 95 % of words. The confidence figure under each result shows how sure the reader was — check low-confidence results, especially numbers.
Can I edit the text before saving it?
Yes. The text appears in boxes you can type in. Correct anything, then copy it or download it as a text or Word file.
Sources
- Tesseract OCR — the open-source text recognition engine
- Tesseract.js — Tesseract running in the browser
- Tesseract documentation — improving the quality of the output
Formulas last reviewed: September 30, 2026 ·How we check our formulas