Drop in a PDF, pick how much to reduce the resolution of the photos inside it, get a smaller file back.
Click to add a PDF or drop it here
Each recognized word sits right where it was found on the page - click a word to edit it, or hover it and click the × to remove it. Click any empty spot on the page to add a new text box there. Check "Keep as image" under a photo to leave it untouched instead of replacing its text.
Add a PDF, then pick one or both options: reduce resolution (choose a percentage) and/or extract text with OCR. With OCR on, click Start to scan the PDF - a page preview appears with every recognized word sitting right where it was found; click any word to edit it, or check "Keep as image" under a photo to leave it untouched. Paginate through every page that had photos, then click Generate to write the edits back as real text; checking "Reduce resolution" too also shrinks the remaining photos before download.
No - only embedded photos (raster images) are resized. Text and vector drawing aren't resolution-dependent in the first place, so there's nothing for this tool to shrink there.
Each page is rendered in your browser, and the PDF's own drawing instructions tell us exactly where each embedded photo sits on the page. The OCR engine reports a bounding box for every word it reads inside that photo, so each word in the review screen sits at the same spot it was found.
Entirely in your browser, using the same Tesseract OCR engine (compiled to WebAssembly) the site's other OCR tools use. The file is uploaded to the server only if you also choose to reduce resolution, as a separate step after you've reviewed and generated the text.
It's left exactly as it was - none of the words recognized from it get written back, and the photo itself is untouched (aside from resolution reduction, if that option is also on).
9 visits