How to OCR a scanned PDF in your browser
Make a scanned PDF searchable and selectable with on-device OCR. Text stays on your machine.
UnboundPDF is a free suite of 53 PDF and image tools that run entirely in your browser — merge, split, compress, edit text, OCR in 126 languages, redact, sign, convert and archive to PDF/A. Your document is read and written by the page on your own device; there is no document-upload endpoint in the core tools, no account, no watermark and no daily cap. Every result can be checked — with the Network tab, or with the Document Passport the Workspace writes for a chain of steps.
- Open OCR PDF and drop a scanned document.
- Run recognition — Tesseract runs in your browser.
- Download the PDF with an invisible searchable text layer over the original pages.
- Optionally open the result in the PDF Editor to fix wording.
Quality depends on scan resolution and clarity. Crooked scans improve after Deskew.
What runs where
The OCR engine is a vendored copy of Tesseract that runs inside your browser. When you drop a scan, the page images are rendered and recognised in your browser process — they are not sent to a remote OCR API. The recognised text is laid over each page as an invisible, selectable layer, so the result looks identical to the original scan but can be searched and copied.
Getting a good result
Recognition quality tracks the scan, not the tool. What helps, in order of impact: scan at 300 DPI or more; keep pages straight — run Deskew PDF first if the text is tilted (a page that is sideways rather than tilted needs rotating instead); prefer clean black-on-white pages over photographs of pages; and pick the right language before running recognition. The picker covers 126 languages grouped by script, with auto-detect if you are not sure — see OCR a PDF in more than 100 languages for the full picker and measured per-language accuracy.
What OCR here does not do
It does not correct wording, and it will not read handwriting reliably — that is true of Tesseract generally, not just this build. If a word comes out wrong, open the result in the PDF Editor and fix it in place.
Why on-device OCR matters for scans specifically
Scans are usually the sensitive documents — IDs, contracts, medical letters. Sending them to a server just to make them searchable is exactly backwards. Verify the no-upload behaviour yourself with the Network-tab method in the PDF Privacy Lab.
Frequently asked questions
Does the OCR here send my scanned document to a server?
No — the OCR engine is a vendored copy of Tesseract that runs inside your browser. When you drop a scan, the page images are rendered and recognised in your browser process; they are not sent to a remote OCR API.