Image to Text
Extract text from any image or pasted screenshot — runs 100% in your browser.
Drop image here or click to select
JPG, PNG, WebP — or paste screenshot (Ctrl+V)
Note: clear, high-contrast images work best. The first recognition downloads a ~2 MB language model.
Extract text from an image using OCR
Upload a photo of a document, a screenshot of a tweet, a scanned page, or a photo of a signboard. The tool runs Tesseract.js, an Optical Character Recognition (OCR) engine, entirely inside your browser via WebAssembly, to pull out the text, which you can then copy and paste. There's no backend at all, your image is never uploaded anywhere. It works best with clear, high-contrast text on a plain background. Handwriting is hit-or-miss, but printed text in English is fairly reliable. The language dropdown only offers five options, English, Hindi, Spanish, French, and German, so other languages aren't selectable at all.
Indian paperwork that needs OCR
- Scanned PDFs of government forms, rent agreements, or mark sheets that are images, not searchable text. Extract the text to make it editable.
- Screenshots of WhatsApp messages or tweets that you want to quote or translate. Instead of retyping, let the OCR pull the words.
- Business cards: snap a photo, extract the name, phone, and email to add to contacts.
Tip: For best results, ensure the photo is well-lit and aligned straight. If the text is tilted, the OCR may misinterpret characters. Use the rotate tool first to straighten it.
Frequently Asked Questions
Can it read Hindi or other Indian languages?
Hindi is one of the five languages in the dropdown (English, Hindi, Spanish, French, German) and works reasonably well on clear printed text. Other Indian languages like Tamil, Telugu, or Bengali aren't in the language list at all, so this tool can't recognize them.
Does it work on handwritten notes?
Rarely, unless your handwriting is exceptionally neat and block-lettered. Cursive and stylised handwriting will not be recognised accurately.
Does it upload my image anywhere?
No. Tesseract.js runs the whole OCR process locally in your browser via WebAssembly; there's no server involved.
Why does the first scan take longer?
The selected language's OCR model has to download the first time you use it, subsequent scans in the same language are faster since it's already loaded.
Can I select multiple languages at once?
No, the dropdown is single-select, choose the one language that matches your image before scanning.