Running PP-OCRv6 in the browser without uploading the image

I recently added OCR to Imging. The main requirement was simple: screenshots, invoices, contracts, and admin-panel exports should not have to leave the device just to extract text. The recognizer runs