How to Extract Text From a Scanned PDF (OCR)
What OCR is, the free options that run on your own device, and why scanned contracts and medical records shouldn't go through someone else's server.
You have a scanned PDF or a photo of a document and you need the text out of it. You could retype it. The alternative is OCR — Optical Character Recognition — which reads the shapes in an image and turns them back into characters you can search, copy and edit.
What OCR actually does
A scanned page is a picture. There are no letters in the file, only pixels arranged to look like letters. OCR analyses those pixels for patterns matching characters, then reconstructs the text. Modern engines use machine learning and clear above 99% accuracy on clean printed material — good enough that the bottleneck is now image quality, not the algorithm.
Common reasons to run OCR
- Scanned paperwork: turning contracts, invoices and receipts into editable, searchable text.
- Phone photos of documents: extracting text from a picture you took of a page.
- Image-only PDFs: files produced by a scanner with no selectable text layer.
- Digitising archives: converting physical records into something a search box can reach.
Free OCR options that run on your device
- iPhone, iPad and Mac: Live Text recognises text directly in the Camera and Photos apps — processed on the device, nothing uploaded.
- Android: Google Lens pulls text out of photos and screenshots instantly.
- Tesseract: the long-standing open-source OCR engine. Command-line, but entirely local and scriptable for bulk jobs.
- Google Drive: upload a scanned PDF and open it with Google Docs and the text is extracted automatically. Free and effective — but the document is processed on Google's servers, which is the trade-off this article is about.
The workflow is the same everywhere: load the document, pick the language, wait for processing, copy the text out.
Docuboxer is building its own OCR that will run in the browser like the rest of the suite, with nothing uploaded. You can follow it on the roadmap.
Why OCR is a privacy decision
Think about what people actually run OCR on: legal contracts, medical records, bank statements, identity documents. These are among the most sensitive files anyone owns, and OCR requires handing the tool the complete, unredacted page.
When processing happens on your own device, the document never reaches a server. That matters especially for:
- Documents covered by GDPR or HIPAA.
- Confidential corporate material.
- Anything containing customers' personal data.
Getting better accuracy
- Image quality: sharp, evenly lit, unskewed pages beat everything else you can do.
- Resolution: aim for 300 DPI when scanning; below 150 accuracy drops sharply.
- Language: select the right one — accented characters get mangled otherwise.
- Text type: printed text is far more reliable than handwriting, which is still genuinely hard.
If the scan is a multi-page PDF, it often helps to split out just the pages you need first, or to export pages as images so you can feed them to an OCR app individually.
Frequently asked questions
Why can't I select the text in my PDF?
Because the page is an image, not text. Scanners and phone cameras produce a picture of the document; there are no characters in the file for the reader to select. OCR is what turns those pixels back into characters.
Is there a free OCR that doesn't upload my document?
Yes, several. Apple's Live Text (iPhone, iPad, Mac) and Google Lens on Android both run on the device. Tesseract is a free open-source engine you can run locally from the command line. Google Drive is free too, but the document is processed on Google's servers.
How accurate is OCR?
Modern engines exceed 99% on clean, printed text at reasonable resolution. Accuracy falls with blur, low resolution, skewed pages, unusual fonts and — most of all — handwriting, which remains substantially harder than print.
Does Docuboxer have OCR yet?
Not yet. It's on the roadmap and will run in the browser like the rest of the suite. Until then, the on-device options listed in this article are the ones that keep your documents private.
Work on your PDFs without uploading them
Merge, split, compress and unlock PDFs 100% in your browser.
See PDF tools →Related tools
- Compress PDF — Shrink scanned PDFs, which are usually heavy.
- Merge PDF — Combine several scans into one file.
- PDF to JPG — Export pages as images for an OCR app.
- Unlock PDF — Remove a password you already know.
You might also like: how to compress a PDF without losing quality and converting images to PDF.