Optimize PDF Tools
Compress and OCR your PDFs to make them smaller and searchable.
Two different problems get called 'fixing a PDF'. One is that the file is too big to email or upload, almost always because it contains full-resolution scans of pages. The other is that the file is a picture of text rather than text — you can see the words but you cannot select, search or copy them, because a scanner produced an image and wrapped it in a PDF. Compression solves the first; OCR solves the second.
Which problem do you have?
- Compress PDF
- The file is too large for an email attachment, an upload form or a portal limit.
- OCR PDF
- You can see text in the document but cannot select it, search it or copy it — it is a scan.
Where a PDF's size actually comes from
Text is almost free: a hundred pages of plain text is a small file. Weight comes from embedded images, and specifically from scans. A page scanned at 600 DPI in colour is a photograph several thousand pixels across, and twenty of those make a document of tens of megabytes. Compression works by re-encoding those embedded images at a resolution and quality appropriate to a document rather than a photo print — which is why a text-heavy PDF barely shrinks and a scanned one can drop by 70% or more. If your file is mostly text and still large, the cause is usually embedded fonts or a handful of oversized images rather than the document as a whole.
What OCR adds, and what it does not change
Running OCR does not replace your scan with typed text. It renders each page, recognises the shapes on it, and adds an invisible text layer positioned behind the image. The page looks exactly the same, but the words are now selectable, searchable and copyable, and screen readers can read them. That layer is a best guess: clean, high-contrast, straight scans are recognised accurately, while faint photocopies, skewed pages, unusual fonts and handwriting produce errors. Because the original image stays on top, an OCR mistake never changes what a reader sees — it only affects what a search finds.
OCR first, then compress
Order matters. OCR reads the page image, so it works best on the sharpest version you have. Compressing first throws away exactly the fine detail the recogniser depends on, and character accuracy drops noticeably. Run OCR on the original, then compress the result — the text layer is tiny and adds almost nothing to the file size.
How much compression is too much
There is a point where a scanned page stops being reliably legible, and it arrives sooner for documents than for photographs, because text lives in exactly the sharp edges lossy compression damages first. If a document will be printed, signed, or read as evidence of something, check the output at full zoom before sending it. If it is a receipt going into an expense system, compress hard. And if a document must stay pristine, resizing the source scan to 200-300 DPI before it ever becomes a PDF beats compressing afterwards.
Frequently asked questions
How much smaller will my PDF get?
Scanned documents commonly drop 50-80%. A PDF that is mostly real text may barely change, because there was little to compress in the first place.
Which languages does the OCR support?
Tesseract ships language models for over a hundred languages, and you pick the one matching your document before running it. Choosing the right language matters — recognising Romanian text with the English model produces markedly worse results.
Why is OCR slow?
Each page is rendered and then analysed on your own CPU rather than on a server. A long document takes minutes. In exchange, the file never leaves your device.
Can I search a PDF after compressing it?
If it already had a text layer, yes — compression re-encodes images, not text. If it was a scan with no text layer, it was never searchable, and compressing it does not change that. Run OCR first.
Is my document uploaded?
No. Both compression and OCR run inside your browser, so the file stays on your machine.