OCR Tool — Extract Text from a Scanned Image or PDF
A scanned document or photo of text has no selectable text layer — OCR (Optical Character Recognition) reads the characters out of the image itself. Upload a photo, screenshot, or scanned PDF page and get back copyable, searchable text — processed locally in your browser using Tesseract.js.
OCR accuracy depends heavily on image quality. The OCR engine (Tesseract.js) loads from a public CDN; your document content is processed entirely on your device.
How to use this tool
- Drop or choose a photo, screenshot, or scanned PDF containing text.
- Wait while the OCR engine loads (first time only) and reads the text — a progress bar shows the status.
- Review the extracted text in the box — copy it or download as .txt.
- Works best with clear, well-lit, reasonably high-resolution scans — blurry or tiny text reduces accuracy.
Frequently asked questions
Is my image or PDF uploaded anywhere?
Your file itself is never uploaded — OCR processing runs locally in your browser using Tesseract.js. The OCR engine's code and language data load once from a public CDN (like any JavaScript library), but your actual document content stays on your device throughout.
Why is the first scan slow?
The OCR engine (a few megabytes of code and trained-language data) has to load the first time you use it in a session. Subsequent pages in the same session are much faster since it's already loaded.
How accurate is browser-based OCR?
Very good for clear, well-lit, reasonably high-contrast text (typed or printed) — often 95%+ accurate. Handwriting, low-resolution photos, unusual fonts or skewed/rotated scans reduce accuracy noticeably. Straighten and crop your image beforehand for the best results.
Can it read languages other than English?
This version is configured for English text. Multi-language support is possible with Tesseract.js but adds significant load time per language — let us know via Contact if you need a specific language and we'll consider adding it.