Extract text from any image — right in your browser

OCR that never uploads. Drop a screenshot, photo or scan — or paste one with Ctrl+V.

OCR that runs on your own machine

Optical character recognition used to mean a server: upload the image, wait, get text back. That is a poor trade for the things people actually run OCR on — a screenshot of a private message, a photo of an invoice, a scan of a passport or a payslip.

TextFit runs a full OCR engine compiled to WebAssembly inside the browser tab. The image is read into memory, recognised locally, and the text appears alongside it. Nothing is transmitted, so nothing needs deleting afterwards. Load the page, disconnect from the network, and it still works.

Three ways to get an image in

Getting the best accuracy

OCR is sensitive to input quality in ways that are easy to control. Sharp, high-contrast, straight-on images read far better than angled phone photos of a page. Screenshots are close to ideal, because the text is already crisp pixels rather than a photograph of ink. If a result comes back garbled, a better capture usually fixes it faster than any setting.

Frequently asked questions

Is the image uploaded?

No. Recognition happens on your own device — which is the point, given what people photograph.

How accurate is it?

Very good on clear screenshots and clean scans; less reliable on handwriting, heavy stylisation, or low-light photos taken at an angle.

What does it cost?

One English image at a time is free. A one-time purchase adds batch OCR and more languages.

Step-by-step instructions · Privacy policy

More private, in-browser tools

Every tool in the suite runs entirely on your own device — same approach, different job. Each is also available as a browser extension.