Image to text

Optical character recognition estimates letters from image pixels. A sharp, evenly lit image with straight lines generally works better than a tilted photograph or tiny screenshot. This tool uses an English recognition model and runs it in your browser. The first run downloads the model from this site; larger images and slower phones need more time. It can help with a receipt, printed note or screenshot you cannot select as text.

A practical example

Paste a screenshot of a printed paragraph, extract the words and compare them with the original before quoting them.

Recognition is not transcription certainty. Check names, decimal points, dates and similar characters such as O and 0. Handwriting, Urdu, Arabic and other unsupported scripts are not promised in this version. The text result does not preserve page design.

Keep a separate copy of the source file. Check the downloaded result in the application your recipient will use, especially page order, small text and important numbers. Conversion is not a guarantee that a portal will accept a document.

Are Urdu and handwriting supported?

This release is for printed English. Other scripts and handwriting are not reliably supported.

Why does the first run take longer?

Your browser loads the recognition engine and English model before processing.

Are my files sent to your server?

The processing code runs in this browser. Input files and generated results are not uploaded by the tool. Visiting the page still makes ordinary requests for site assets, and the hosting server may record those requests. Read the tool privacy explanation.