Make a scanned PDF searchable
Make a scanned PDF searchable, without changing how it looks.
Drop your file here
Processed securely and deleted within the hour.
The problem this solves
A scanned PDF looks like a document and behaves like a photograph. You cannot search it, you cannot copy a sentence out of it, and a screen reader finds nothing to read. Every page is just an image of paper.
OCR reads the characters in those images and adds them back as real text — invisibly, drawn exactly over the words they correspond to. The page looks precisely as it did before; it has simply become searchable, selectable and accessible.
Why the page looks unchanged
The original page is kept untouched and the recognised text is drawn on top in an invisible render mode. Nothing is re-rendered, so the scan keeps its exact appearance, its quality and its colours. Select a line in the result and you will see the highlight land on the ink underneath.
This matters for anything official. A re-rendered scan of a certificate or a contract is a different document; an OCR'd one is the same document with a text layer added.
Scan resolution
200 DPI suits most scans and is the default. Push to 300 or 400 for small print, faint type or documents that scanned badly — recognition improves, at the cost of time. Below about 150 DPI accuracy falls away quickly, which is why 120 is the floor here.
Accuracy, and being honest about it
Clean printed text is recognised very reliably. Faint photocopies, skewed pages, handwriting, unusual typefaces and heavy background patterns all reduce accuracy. The result reports the average confidence, and warns you when it is low rather than letting you assume a perfect transcription.
Always keep the original. OCR adds a layer; it does not verify it.
If your PDF already has text
Then it needs no OCR at all — PDF to Text extracts it exactly, with no recognition step and no chance of error. Run that first; if it reports no text layer, come back here.
Common questions
Will my document look different?
No. The original pages are untouched — the recognised text is added invisibly on top.
What resolution should I use?
200 DPI for most scans, 300 or 400 for small or faint print. Higher takes longer.
How accurate is it?
Very good on clean printed text. Faint photocopies, skew and handwriting reduce it. The average confidence is reported, and flagged when low.
Does it read handwriting?
Poorly. These models are trained on printed text.
Which languages work?
Latin scripts and Chinese are handled well by the bundled models. Other scripts vary.
Is my document kept?
No. It is processed in an isolated directory and deleted once the download is sent.