Scanned PDF OCR
Recognize Chinese or English text from scanned PDF pages, review, copy and export TXT.
Instructions and limits
- Upload a scanned PDF, choose pages and Simplified Chinese, Traditional Chinese or English, and adjust orientation if needed.
- Recognize pages in sequence. Stop to keep completed pages, then review, copy or download the text as TXT.
PDFs: 80 MB and 1000 pages; 20 pages per run, 200,000 characters per page and 2 million total. Raster size: a 3000 px edge and 6 MP; embedded images above 24 MP may be skipped. The engine and models download on first use; files stay local. Encrypted PDFs are unsupported. Best for clear printed text; proofreading is needed. Outputs plain text, not searchable PDFs or reconstructed tables and columns.
Use cases and example
Turn scanned PDF pages that cannot be selected or copied into editable text, useful for digitizing printed documents and archived scans.
Upload a scan, select pages 1–3 and the appropriate language, and start OCR. Review names, numbers and line breaks on each page, correct errors and download TXT with page separators.
Processes pages locally, up to 20 per run. Stopping keeps completed pages. Speed and accuracy depend on the device, image clarity and text size. Handwriting and complex tables may be misread. No text layer is added to the PDF; use PDF to text for documents that already have one.