Nightly builds of Firefox have begun testing OCR, which allows you to extract text from images hosted on a web page and place the recognized text on the clipboard or speak to visually impaired people using a speech synthesizer. Recognition is performed by selecting the "Copy Text from Image" item in the context menu displayed when editing with the mouse button on the image.
The feature is currently enabled only on the macOS platform and will also be available in builds for Windows soon. The implementation is tied to the system OCR APIs: VNRecognizeTextRequestRevision2 for macOS and Windows.Media.OCR for Windows. There are no plans to implement the feature for Linux yet.
Source: opennet.ru