Scanned PDF to Clean Text

Select PDF pages, recognize their printed text locally and download cleaned TXT with ordered page markers.

OCR code and language models download by GET. Your image/PDF and recognized text stay in this browser; no document upload. Reset clears this workflow session.

This flow needs JavaScript. If it does not load, refresh the page and try again.

About this flow

Select PDF pages, recognize their printed text locally and download cleaned TXT with ordered page markers. Choose explicit settings, review the real downloaded output and keep your original for retry.

Limitations

Initial local OCR models/code download by GET; PDF contents are never uploaded. Documents: 50 MiB/500 pages; OCR: up to 20 selected pages and 1,048,576 extracted UTF-16 code units. TXT preserves --- Page N --- markers, but omits confidence/boxes and original layout. Recognition varies; check results. Output is text, not a searchable PDF. No cold-start-offline or handwriting accuracy promise.