1. Prepare
Open a PDF and explicitly choose pages, such as 2,1. Select English, Spanish or Portuguese. The example is an image-only PDF, so its text must be genuinely recognized.
PDF OCR to Text
More actions
PDF and recognition selection
Browser: PDF up to 15 MiB and 200 pages; choose up to 3 pages per job. Rendering uses 150 dpi, up to 4 million pixels per page and 9 million total; 60-second job limit.
Use 1-based pages in the desired order, for example 1, 3-4 or 2,1. Select no more than 3 pages; no extra pages are processed silently.
Recognition supports English, Spanish and Portuguese only. The interface language does not change the recognition model. Loading recognition models from this site may require a network connection. Your PDF stays in the browser and is not uploaded.
OCR uses page images, including scanned PDFs; it produces plain text without preserving PDF page layout and adds no searchable PDF layer. Encrypted PDFs, XFA and pages with interactive widgets are refused. Inline images, Type3 fonts and unsupported page-content compression are refused before rendering. Open the PDF editor for page operations.
Found something we could improve?
Practical usage guide
A practical workflow for PDF OCR to Text
Recognize selected PDF pages locally in English, Spanish or Portuguese. Review per-page OCR confidence and export UTF-8 text with page boundaries.
2. Review
Review recognized text and per-page confidence against the original. Check names, numbers, accents and missing words; confidence is not measured accuracy.
3. Apply
Copy or download UTF-8 TXT with page boundaries. Browser jobs allow up to three pages; OS jobs allow one selected page. No searchable PDF or Word layout is produced. Keep the original and use flattened copies of interactive forms. Inline images, Type3 fonts and unsupported page-content compression are refused before rendering.
