Image to Text (OCR)

Text locked inside an image is frustrating: a screenshot of an error message, a photo of a receipt, a scanned letter, a slide from a conference, a sign in a foreign language. Optical character recognition turns those pixels back into editable text. This tool runs OCR directly in your browser and gives you text you can correct, copy or save.

How it works

The page uses Tesseract, a long-established open-source OCR engine, compiled to WebAssembly. The engine and the language data are loaded from this site the first time you use a language and then cached by your browser. Recognition happens on your device, so even confidential documents never leave it.

Choosing languages

Pick the language of the text from English, Spanish, French, German, Portuguese, Italian, Dutch, Hindi, Russian, Indonesian or Turkish. Choosing the right one matters: each language model knows its alphabet, accents and common words, which improves accuracy noticeably. For documents that mix two languages, such as a bilingual form or an English letter with a German address, add a second language.

Getting the image in

Drop a file, choose one, or simply paste a screenshot with Ctrl+V (Cmd+V on a Mac). JPG, PNG, WebP, iPhone HEIC, AVIF, TIFF and other formats are all accepted. Screenshots give the best results because their text is perfectly sharp.

The contrast option

Photos of paper are often unevenly lit, with shadows from your hand or phone and gray rather than white paper. “Improve contrast” evens that out before recognition and is on by default. For light text on a dark background, such as dark-mode screenshots or white lettering on signs, switch it off if the result is empty or garbled.

Reading the result

The recognized text appears in an editor with the average confidence next to it. You can fix mistakes directly, then copy the text or download it as a .txt file. Always double-check numbers, names, email addresses and amounts, because OCR errors in those are easy to miss and costly.

Tips for difficult images

Photograph documents straight on, not at an angle, in good light. Fill the frame with the text. Avoid motion blur by resting the phone on something. Small print needs more pixels, so take the photo closer rather than zooming in later. For a photographed page that is skewed, straighten it first with Perspective Crop Image. Cropping away surrounding clutter with Crop Image also helps.

Everyday uses

Copying an error message from a screenshot into a search engine, pulling an address from a photo of a business card, digitizing a printed recipe, grabbing a quote from a book page, extracting serial numbers from a device label, or turning a slide photo into notes. Translating a sign or menu works well too: recognize the text in its language here, then paste it into a translator.

What OCR handles poorly

Handwriting, heavily stylized fonts, text over busy photos, very low-resolution images and complex table layouts are hard. Text is returned in reading order line by line, so columns and tables may need rearranging afterwards.

Related tools

For whole scanned documents, OCR PDF recognizes every page. To read the link or text stored in a QR code, use the QR Code Scanner. For extracting text from normal PDFs, PDF to Text is exact and instant.