Pull selectable, editable text out of photos and scanned images using on-device OCR | no upload required, and no data ever leaves your browser.
Upload a photo or scanned image containing text.
Shrivo runs an on-device OCR model (Rust + ONNX) to detect and recognize the text.
Copy the extracted text or download it as a plain text file.
Accuracy depends on image clarity | the on-device engine handles printed text, screenshots, and moderately skewed scans well, though very low-resolution or handwritten text may be less reliable.
Once the page and OCR model are loaded, recognition runs fully offline in your browser.
The OCR engine is optimized for Latin-script languages, such as English, out of the box.
No | recognition runs entirely on-device using WebAssembly and ONNX, so your images and extracted text never leave your browser.