OCR a PDF — make a scanned PDF searchable without uploading it
Real OCR in your browser: the engine itself runs on your device.
Free: 3 jobs per UTC day, 25 MB per file. PDF exports include a visible Made with Bindery watermark. Pro exports clean PDFs.
Prefer the full studio? Open OCR PDF in the app →
Bindery ships the Tesseract OCR engine inside the app — about 7 MB that loads once, from Bindery itself, with no cloud service involved. It reads each scanned page and lays the recognized words invisibly on top of the untouched image, so your scan looks identical but becomes selectable and searchable in any PDF viewer. Prefer raw text? Get a .txt instead.
OCR is where privacy matters most — scans are usually contracts, IDs and records. With Bindery the pages are never transmitted anywhere: the recognition happens on your machine, full stop.
How it works
- Drop your scanned PDF and open OCR
- Choose searchable PDF or plain text output
- Download — pages look identical, text is now selectable
How OCR in the browser works
Optical character recognition turns a picture of text into text. Bindery ships the Tesseract engine inside the app — about 7 MB, downloaded once from Bindery itself and then cached — and runs it in your browser. Each page is rendered, analysed, and the recognised words are laid invisibly over the untouched page image at their exact positions. The scan looks identical; it is now selectable, searchable and copyable in any viewer, and screen readers can speak it. Choose plain text output instead when you only want the words.
Six language packs are available — English, Spanish, French, German, Portuguese and Italian — each loaded on demand and cached on your device. Pick the language of the document; a French letter recognised as English loses its accents and many of its words.
Getting accurate results
- Resolution: 300 DPI scans recognise best. Phone photos work when the page fills the frame and is in focus.
- Straight and clean: run Clean scanned PDF first for skewed, faint or speckled pages — deskew and contrast make the biggest difference to accuracy.
- Borders: dark scanner edges confuse layout analysis; crop them away.
- Check names and numbers: OCR is excellent on running text and weakest on handwriting, stylised fonts and tables of digits. Spot-check the parts that matter before relying on them.
- Compress after, not before: OCR the best copy you have, then compress the searchable result.
OCR in the browser versus uploading
| OCR PDF | Bindery | Upload-based tools |
|---|---|---|
| Where the file is processed | In your browser tab, on your device | Uploaded to the service’s servers |
| Engine | Tesseract, running on your device | A cloud OCR service |
| Languages | Six on-device packs, cached | Usually more, server-side |
| Scans of IDs, contracts and records | Never transmitted | Uploaded page by page |
| Free allowance | 5 pages a day; Pro unlimited | Varies |
| What exists afterwards | The original and the new file, both on your device | A server copy until the retention window ends |
| Works offline | Yes, once the app has loaded | No |
What to do with a searchable scan
Once the text layer exists, every text-based tool works: PDF to Word and PDF to text can read it, Compare can diff two scanned versions, Redact can find and remove the text you mark, and Ask your PDF can answer questions about it — all on the device. Accessibility and archive checks (accessibility, PDF/A) also require a text layer on every page.
Questions people ask
Which languages are supported?
English, Spanish, French, German, Portuguese and Italian — each an on-device language pack that loads once. Recognition quality is strongest on clean, straight scans at around 300 DPI; run Clean scanned PDF first if the pages are skewed or faint.
Does OCR change how my scan looks?
No. The searchable-PDF output keeps every page pixel-identical and adds an invisible text layer on top for selecting and searching.
Is my scan sent to a cloud OCR service?
No — this is the rare OCR tool that runs entirely in your browser. The engine is part of the app itself.
How accurate is browser OCR compared with cloud OCR?
On clean, straight scans of printed text at around 300 DPI, Tesseract is highly accurate in the six supported languages. Cloud services are typically better on handwriting, low-quality photos and exotic layouts. For contracts, statements and letters, the on-device result is usually good enough to search, copy and index — check names and figures by eye.
Does OCR work on a photo of a document?
Yes, if the page is in focus, fills the frame and is reasonably flat. Scan from phone straightens pages as you capture them; Clean scanned PDF fixes skew and contrast on existing photos.
Is there a page limit?
The free plan covers 5 OCR pages a day and files up to 25 MB; Pro removes both limits. Beyond that, your device’s memory is the only practical limit.