HI-RES OCR
Detection → deterministic reading-order reconstruction (deskew + column split + line clustering) → recognition.
Pick the tab that matches your input — handwriting vs printed text (not language): printed English belongs in the multilingual tab.
TrOCR-large + HI-RES pipeline — English only. Runs on free CPU, so a full page may take ~20–60s. Please be patient.
Recognized text