GramSeva Mitra
← Smart Writing Studio

Browser OCR Text Extractor

Next-gen client-side OCR — no Tesseract, no server. Model downloads once and caches in your browser.

Image input

Powered by Transformers.js + ONNX Runtime WASM (TrOCR). Preprocessing runs on Canvas before inference in a Web Worker.

Click “Load OCR model” to download (~150 MB, cached after first use)

Extracted text