Extract text from PDFs
Read a PDF’s text layer directly and use OCR on pages without a text layer.
Add a PDF and choose a page range. Each page is checked for a text layer: when present, text is extracted without OCR; otherwise the page is rendered as an image for recognition.
Leave the range empty to process every page. Combine ranges and individual pages, such as 1-3,5. Page numbers start at 1 and must be within the document’s page count.
Edit page results and copy or save all pages or the selected page as TXT. Text layer order and font information can affect spacing and reading order. The tool does not reconstruct table structure.
How to use
- Select a PDF and check its page count.
- Optionally enter a range such as 1-3,5 and start extraction.
- Review each page and choose which results to copy.
What to check
Your selected OCR mode is not applied to pages with a text layer. There is currently no option to ignore that layer and force OCR.
Frequently asked questions
Do text PDFs need a model?
Pages with a text layer are extracted directly using pdfjs-dist. Pages without one require an OCR engine and the selected language data.
Can I leave the page range empty?
Yes. An empty range means all pages. An invalid range is reported before processing starts.
Can a PDF mix scanned and text pages?
The text layer is checked separately for each page. Only pages without one go through OCR.