OCR PDF
OCR PDF is a free tool that runs optical character recognition on scanned PDF pages directly in your browser. Pick your languages, get the text as TXT or a searchable PDF. The language models download once and are cached; your documents never leave the device.
Drop PDF files here or click to browse
Files stay in your burrow: everything is processed on your device, never uploaded.
How to use this tool
- 1Drop a scanned PDF into the tool.
- 2Pick one or more recognition languages and the scan resolution (150 or 300 DPI).
- 3Choose the output: plain text, a searchable PDF (page images plus invisible text layer), or both.
- 4Click Recognize and watch the per-page progress; download the result when it finishes.
FAQ
Why does the first run take so long to start?
The recognition language model has to be downloaded once, about 1 to 3 MB per language. After that first download it is cached on your device and later runs start immediately.
Does OCR work on every scanned PDF?
No, scan quality decides. Straight, even pages scanned at 200 DPI or more recognize well. Skewed pages, dim or uneven lighting, handwriting and heavy noise all reduce accuracy. For photos of pages taken with a phone, expect to fix errors manually afterwards.
How accurate is this compared to server-side OCR services?
It is honest, good OCR, not server-grade precision. Clean printed text in a supported language usually comes out highly accurate; tables, multi-column layouts, formulas and mixed scripts can produce errors. The searchable PDF output keeps the page as an image with an invisible text layer, so search and copy work but the look never degrades.