Feedback

Recognition runs on your device. Sharing a correction is different: once you tick the consent box and share, the corrected text, its error type and the file you attached (or the original output, if you attached none) are uploaded to the MonOCR project for model training.