Improve Smart Selection and OCR Language Support

Please improve the Smart Selection and OCR (text extraction from images) feature.

Since the latest update, the text extraction feature has become unstable. Sometimes it works correctly, while at other times it fails to recognize or extract text from the same type of image.

Please investigate and fix the reliability of the OCR system so that text extraction works consistently.

There is also an unnecessary limitation where users can only use one OCR language for a period of 15 days. This is not practical, especially for users who regularly work with multiple languages.

OCR should automatically support multiple languages without requiring the user to wait 15 days to change the selected language.

Ideally, the system should automatically detect the language of the text in the image and use the appropriate OCR model. It should also support mixed-language images, such as Arabic and English appearing together in the same document or image.

Users should be able to extract text from any supported language whenever needed, without a 15-day language restriction.