Image to Text (OCR)

Optical character recognition, OCR for short, turns the letters in a picture into text you can copy, edit and search. It saves you from retyping text you can see but can't select: a paragraph in a screenshot, a quote from a book page, a recipe card, an address on a letter, a serial number on a label, notes on a whiteboard, the details on a receipt for an expense report, or the text of a scanned document.

The result can be pasted straight into an email, document or spreadsheet, or downloaded as a .txt file.

Students use it to copy quotes from textbook photos, and office workers use it to turn printed invoices and letters into text they can file and search.

Accuracy depends mostly on the image. Photograph the text straight on rather than at an angle, in even light without shadows or glare, and hold the phone steady so the letters are sharp. Fill the frame with the text, and crop away everything else, such as the edge of the table or neighbouring pages; Crop Image helps with that.

Clean, printed text at a normal size is usually recognised with over 95% accuracy. Small images are enlarged automatically before recognition, which helps with screenshots of small text. For a single paragraph or column, the One block of text layout often gives cleaner results than automatic layout.

OCR works best on printed text. Handwriting, decorative fonts, text on curved surfaces like bottles, very low resolution images and text over busy backgrounds produce more mistakes, so check important results such as names, numbers and amounts. The tool shows a confidence score to help you judge. It reads English; other languages using the Latin alphabet may partly work, but accented letters can come out wrong.

Tables are read as lines of text, so columns may need tidying after pasting into a spreadsheet.

It reads JPG, PNG, WebP, HEIC and other image formats, so iPhone photos work directly, and you can process several images in one go. The first run downloads the OCR engine once, about 7 MB, and later runs start immediately.

If your text is inside a PDF, try PDF to text first: PDFs made from documents already contain real text that can be copied without OCR. For scanned PDFs that contain only images of pages, use the scanned PDF to text tool, which reads every page with the same engine.