A photographed receipt
A slightly tilted phone photo of a printed receipt
Most of the text recognized correctly; rotate the image first if it's sideways, since that alone often fixes the largest source of errors.
Image to Text (OCR)
Extract editable text from a photo or screenshot in English or Turkish, entirely in your browser.
The first run for a language downloads its data (a few MB) and caches it — later runs start immediately.
Extracted text (editable)
The result will appear here.
Was this tool useful?
About this tool
OCR (optical character recognition) turns the pixels of a photographed or scanned page back into actual, selectable text — the same technology behind a scanner's "searchable PDF" feature, running here entirely on-device via Tesseract, an open-source engine originally developed at HP and now maintained by Google.
How to use it
- Upload a photo or screenshot containing text, and pick the language — English, Turkish, or both if the image mixes them.
- The first run downloads that language's data (a few megabytes, cached afterward), which shows as a loading stage before recognition starts.
- Once done, the extracted text appears in an editable box — fix anything OCR got wrong, then copy or download it as .txt.
Examples
Common problems
- The recognized text is full of errors.
- OCR accuracy depends heavily on image quality — low resolution, blur, skewed angles, and low contrast between text and background all hurt it. Straighten the rotation, and where possible use a higher-resolution source image.
- The first run takes a long time before anything happens.
- That's the language model downloading — a few megabytes for English or Turkish. It's cached after the first successful run, so later recognitions on the same browser start immediately.