Recognition runs on your machine; nothing is transmitted, and the tool keeps working offline once the app has loaded. Quality follows the scan: straight, clean pages recognize well, while skewed or faded ones benefit from Clean scanned PDF first (before OCR, since cleanup rebuilds pages as images). Choose plain-text output for a .txt with a marker per page. OCR is not part of the bindery-pdf CLI, so for scripted recognition on a server ocrmypdf remains the right tool; for a document on your desk, the browser is faster to reach. Spot-check names and figures against the page.
How it works
- Open Bindery in Firefox or Chrome and drop the scan into OCR
- Choose the language and searchable-PDF or plain-text output
- Download and test with Ctrl+F in your viewer
Questions people ask
Is it the same engine as the tesseract package?
Yes — Tesseract, compiled to run in the browser, with its standard language packs. Results on clean scans match what the command-line tool produces; the difference is that nothing is installed and the page images are handled by Bindery’s own renderer rather than Ghostscript.
Can I run it from a script?
Not through bindery-pdf, whose commands cover page operations rather than recognition. For batch OCR on a server, ocrmypdf is the tool to script. Bindery’s OCR is for the browser, where it runs in a background worker and can be cancelled.
Which languages are available?
Eleven: English, Spanish, French, German, Portuguese, Italian, Arabic, Simplified Chinese, Japanese, Russian and Hindi. Choose the document’s language before running; the engine uses the pack to resolve ambiguous shapes, and each pack loads once from Bindery the first time it is used.