How to extract text from a PDF
The page list marks each page as Text layer detected or Image-based, and the default queue contains image-based pages. Choose Language(s), Rotation for OCR, and Page range, then export text or a searchable PDF. The PDF is processed in your browser and is not uploaded by this tool.
- Drop a PDF file into the drop zone, or click
browse file. - Set
Language(s),Rotation for OCR, andPage rangeunderOCR options. - Choose a mode under
Which pages to OCR, then adjust the page checkboxes if needed. - Review
Review pages, including the Text layer detected or Image-based label and existing character count. - Click
Extract, check the results table, then chooseDownload extracted text (.txt)orDownload searchable PDF.
How to OCR a PDF page by page
Only image-based pages (recommended) checks pages whose extracted text is empty. Force OCR on every page in range checks every page in the range, including text-layer pages. Use my manual selection below only keeps the current checkbox choices. Use Select image-based pages, Select all in range, or Clear selection to change the queue.
Page range: 1-3,7,10-12
Pages in range: 1, 2, 3, 7, 10, 11, 12Extract text from a scanned PDF with language and rotation settings
Language(s) starts with English and accepts multiple selections. The selected languages apply to each page queued for OCR. Rotation for OCR can leave a page unchanged or apply a 90, 180, or 270 degree clockwise rotation. This affects OCR reading and the page image used in the searchable PDF.
Page range accepts comma-separated page numbers and ranges such as 1-3,7,10-12. Leave it blank to include every page. Invalid tokens add no pages, and range endpoints stay within the file's page count. OCR and extracted text use the selected range. Pages outside it stay unchanged in the searchable-PDF output.
How to turn a scanned PDF into a searchable PDF
The results table lists Page, Source, Confidence, Chars, and Status. Pages kept from an existing text layer show 100% confidence. Skipped pages show no confidence value, and OCR failures include an error message. The text download covers the selected range and labels each page as OCR, text layer, skipped, or error.
The searchable-PDF export keeps the source page count and order. OCR pages are rendered as JPEG images with an invisible OCR text layer, so their appearance can change after rotation. Pages that are not OCR'd are copied from the source PDF. A recognized character that cannot be placed in the embedded font is omitted from the invisible layer while the page image stays intact.
Use a page range for part of a PDF
Set Page range when only part of the file needs processing. The page list shows pages outside the range as disabled, and the text export includes only pages inside it. Leave the field blank to process the whole file.
Use Rotation for OCR for a sideways or upside-down scan, then check the confidence and status values in the results table. The searchable PDF keeps pages outside the OCR selection from the source PDF.
When to use a different PDF extraction tool
Use PDF Extract Tools when choosing another PDF extraction job. Use PDF & Image Table Extractor when the output needs table rows, columns, headers, or merged cells. Extract Text From PDF reads running text and does not rebuild table structure.
One PDF is accepted at a time. Files with more than 300 pages and password-protected or encrypted files are refused during loading. The PDF is processed in the browser, and the code does not upload it.
Questions about extracting text from a PDF
Can I extract text from a scanned PDF?
Yes. Image-based pages are selected for OCR by default. Choose the language, rotation, and pages to process before downloading the text or searchable PDF.
Does Extract Text From PDF OCR every page?
The default mode queues image-based pages only. Choose Force OCR on every page in range to queue checked text-layer pages too.
How do I extract text from selected PDF pages?
Enter page numbers or ranges in Page range, adjust the checkboxes in Review pages, and click Extract. The text download contains only pages inside the selected range.
Will the searchable PDF look the same as the original?
OCR-selected pages can look different because they are rendered again as JPEG images, especially when rotation is applied. Pages that are not selected for OCR are copied from the source PDF.
Can I open a password-protected PDF here?
Password-protected or encrypted PDFs are refused during loading. The tool does not attempt to bypass that protection.