image to text (OCR)
Screenshot, photo or scanned PDF in — editable text out. 15 languages including Hindi, Tamil, Bengali and Urdu.
First use downloads about 6.9 MB (OCR engine + language model). After that it's cached and works offline.
How to copy text from an image
- Pick the language(s) in your image — select more than one for mixed text, e.g. Hindi + English.
- Drop the image or PDF, tap to choose it, or just press Ctrl+V to paste a screenshot.
- Wait a few seconds while it's read on your device. The first run downloads the OCR engine once.
- Fix any mistakes in the text box, then copy it or download it as a .txt file.
Get the most accurate results
- Screenshots winText captured straight from a screen is perfectly sharp — the best possible input.
- Shoot straightFor paper, hold the phone parallel to the page in good light, with no shadows.
- Crop itCut away borders, photos and logos so only the text you need is left.
- Right languageSelecting the actual language matters more than anything else for non-English text.
OCR by language
- English image to text
- Hindi image to text
- Marathi image to text
- Tamil image to text
- Telugu image to text
- Bengali image to text
- Gujarati image to text
- Kannada image to text
- Malayalam image to text
- Punjabi (Gurmukhi) image to text
- Urdu image to text
- Arabic image to text
- French image to text
- Spanish image to text
- German image to text
What works and what doesn't
This tool uses Tesseract, the open-source OCR engine originally developed at HP and later Google, running locally through WebAssembly. It's excellent at printed text in standard fonts, screenshots, documents, book pages and receipts. It is not built for cursive handwriting, text on curved surfaces, or very stylised logo lettering. The result is plain text — bold, colours and table layouts aren't kept.
Need the opposite — turning text into speech? Try text to speech. Want to tidy the extracted text? Use remove line breaks or the word counter.
Frequently asked questions
Is this OCR really free?
Yes — no sign-up, no page limits and no watermark. The OCR engine runs inside your browser, so there's no server cost to pass on to you.
Are my images uploaded?
No. Your images and PDFs are read on your own device. The only downloads are the OCR engine and language models, fetched once and cached by your browser.
Which languages are supported?
English, Hindi, Marathi, Tamil, Telugu, Bengali, Gujarati, Kannada, Malayalam, Punjabi (Gurmukhi), Urdu, Arabic, French, Spanish, German. You can select several at once for mixed-language documents.
Can it read handwriting?
Only very neat printed handwriting, and with mistakes. It's designed for printed and on-screen text.
Can I convert a scanned PDF to text?
Yes. Drop the PDF and each page is rendered and read in your browser, up to 30 pages at a time.
How accurate is it?
On clean screenshots and printed pages, usually very accurate. Blurry photos, unusual fonts and complex scripts (like Urdu Nastaliq) produce more errors. We show a confidence score so you know when to proofread.