OCR — Image to Text
The OCR engine and language data download once on first use, then run entirely on your device — images are never uploaded. Works best on printed text and clear photos; handwriting is not well supported.
About the OCR Tool
How to use it
1. Pick the text language — English, Bangla, or both. 2. Upload an image, screenshot or scan (keep enhancement on for best accuracy). 3. Click Extract Text and wait for the result. 4. Fix any small mistakes directly in the text box, then copy or download the .txt file.
How accurate is it?
Very accurate on printed text, documents and screenshots — typically above 90%, with a confidence score shown with every result. Photos of documents work well when they are sharp and well-lit. Handwriting and heavily stylized fonts are the weak spots.
How enhancement improves results
The optional enhancement upscales small documents 2–3x, converts the image to grayscale and stretches the contrast to the 1st–99th percentile. This removes shadows and camera noise that confuse the OCR engine — a big accuracy boost on phone photos. Turn it off for already-clean black-and-white scans.
Privacy
The OCR engine runs entirely on your device — your images are never uploaded. The engine and language data download once from a CDN on first use, then work from your browser's cache.