The image to text (ocr) works like this: it runs OCR (Tesseract WebAssembly) entirely on-device to extract text from images and screenshots. Everything happens on your device - the input never leaves your tab unless the tool clearly states it performs a server-side lookup.
How to use the Image to Text (OCR)
Select or drop your image.
Adjust the available settings while watching the preview.
Check dimensions, quality and format before exporting.
Download the result and keep the original when quality matters.
Features
Runs on your device via the Canvas and WebAssembly APIs
Batch-friendly when you have several files
Original images stay untouched
Common mistakes to avoid
Over-compressing images that contain small text
Exporting transparency to JPEG
Examples
Input
Result
Photo of a printed page
Editable text output, 20+ languages selectable
Best used for
Getting a reliable answer for image to text (ocr) without installing an app
One-off tasks done in under a minute from any device
Checking results before committing them to something important
Working on a phone where desktop software is unavailable
Privacy
Images are processed on your device; nothing is uploaded unless a tool explicitly says a server lookup is involved.
Frequently asked questions
How accurate is it?
Clean printed text reaches 95%+; handwriting, low contrast and skew hurt. Straighten and crop first for best results.
How accurate is the OCR?
Sharp, well-lit, straight text extracts nearly perfectly; stylised fonts, low contrast and skewed photos degrade it - rotate and crop first when a scan misbehaves.
Does it recognise handwriting?
Print-style handwriting extracts reasonably; cursive is unreliable - OCR is built around typeset text.