Image to Text (OCR)

Extract text from images using optical character recognition.

Image

Extracted Text

About the Image to Text (OCR)

The Image to Text tool runs optical character recognition on a picture and returns the words it finds as selectable, copyable text. Upload a scan or a photo of a document, choose the language, and it extracts the content.

Recognition is done by Tesseract, the open-source OCR engine, compiled to run in the browser. That means the image is not uploaded to a server, which matters for anything confidential, but it also means the first run downloads a language model and takes a few seconds.

Results depend heavily on the input. A flat, evenly lit scan at 300 dpi is close to perfect; a phone photo at an angle, with shadow or a patterned background, will produce errors. Cropping to the text and straightening the page before uploading improves accuracy more than any setting.

Common questions about Image to Text (OCR)

What is OCR?

Optical character recognition: software that reads the shapes of letters in an image and converts them into machine-readable text.

Which languages are supported?

The major Latin-script languages plus a few others. English gives the best results.

How accurate is it?

On a clean 300 dpi scan, close to perfect. On a skewed phone photo, expect mistakes with similar characters and small text.

Is my image uploaded?

No. Tesseract runs locally in the browser, though the engine itself is downloaded from a CDN on first use.

Does it preserve formatting?

Partly. Line breaks usually survive; columns, tables and fonts do not.

Why is the first run slow?

The language model has to download before recognition starts. Later runs are faster.

Related Tools in Everyday Utilities

Password GeneratorQR Code GeneratorRandom Number Generator