About Image to Text (OCR)
Image to Text turns pictures of text into editable, copyable text. Grab the words from a screenshot, a photo of a sign, a scanned letter or a page of a book, without retyping a single line.
How it works
This is OCR (Optical Character Recognition) powered by Tesseract, the open-source engine originally developed at HP and later maintained by Google. It first finds the blocks and lines of text in the image, then recognizes each word with an LSTM neural network trained for the language you select. The language data is downloaded once and cached, so later scans start immediately.
The optional Enhance image step converts the picture to grayscale, boosts contrast and enlarges small images before recognition. This often helps a lot with phone photos of documents.
What you can use it for
- Copy text from screenshots of apps, videos, PDFs or websites that block selection.
- Digitize paper documents: letters, receipts, forms and printed notes.
- Quote books and articles from a photo instead of typing them out.
- Translation: extract foreign text from photos, then paste it into a translator.
Tips for the best results
- Choose the correct language. Recognition is much more accurate with the right dictionary and character set.
- Photograph documents flat, straight on and in good light. Shadows and angles are the main causes of errors.
- Text should be at least 20 px tall in the image. Zoom in before taking a screenshot of small text.
- For PDFs, export each page as an image and drop them all at once. They're processed in order.
- Handwriting and decorative fonts are hard for any OCR engine, so proofread the result.