What the Image to Text Converter (OCR) does
Retyping text from a photo or a scanned page is slow and error-prone. Optical character recognition (OCR) does it for you: it finds the letters in an image and turns them into editable text. With this free image to text converter you can extract the text from a photographed letter, a scanned contract, a screenshot of a chat or an error message, a slide, a receipt or a page from a book.
Choose the language of the text – English, German, Spanish or Persian, or one of them together with English – and the tool shows the recognized text with the number of words and a confidence score. You can copy the result with one click or download it as a plain text file. Images can be selected, dragged onto the page or pasted straight from the clipboard, which is the fastest way to convert a screenshot to text.
Unlike most online OCR services, this tool does not send your images anywhere. The open-source Tesseract engine runs inside your browser using WebAssembly, so even confidential documents such as IDs, bank statements or medical letters stay on your device. There is no sign-up, no page limit and no watermark.
When to use it
Documents and letters
Turn a photographed or scanned letter, form or contract into text you can edit, search or translate.
Screenshots
Copy text from a screenshot of a chat, an error message, a video frame or an app that does not allow copying.
Study and research
Extract quotes from book pages, slides and printed handouts instead of typing them by hand.
Receipts and labels
Capture amounts, addresses, serial numbers or ingredients from receipts, packaging and signs.
Tips for the best results
- Crop the image to the text area first – borders, photos and backgrounds can confuse the recognition.
- Take photos straight from above in good light; slanted or curved pages reduce accuracy.
- For very small text, use a higher-resolution image or zoom in before taking the screenshot.
- Check numbers, names and punctuation in the result – OCR may confuse similar characters such as 0 and O or 1 and l.
How the Image to Text Converter (OCR) works
When you start a recognition, your browser loads the Tesseract OCR engine compiled to WebAssembly and the trained model for the selected language. Both come from this website and are cached, so the next recognition starts without downloading again.
The image is first drawn onto a canvas: small images are enlarged and very large ones reduced, because the engine reads letters best at a certain size, and transparent areas are filled with white. Tesseract then finds lines and words and uses a neural network (LSTM) to recognize the characters, including right-to-left scripts such as Persian.
The result is cleaned of trailing spaces and extra blank lines and shown with an average confidence score. A low score is a hint that the image was blurry, too small or in another language than the one selected.