फोटो से टेक्स्ट निकालें — हिंदी + English
Pull the text out of any photo, screenshot, scanned page or PDF — Hindi and English, even mixed in one line, with ₹ and digits intact. Edit it, copy it, save it as .txt. The OCR runs on your own GPU, so your bills and documents never leave this tab.
Detecting your GPU…
Yes. One model reads Devanagari and English letters, digits and the ₹ sign together, so mixed lines like "कुल राशि ₹12,500" or "PAN card की प्रति" come out right. Marathi, Nepali and Sanskrit printed in Devanagari work too.
No. The OCR model downloads once (about 13 MB) and runs inside your browser on your GPU, or CPU if there is no GPU. Bills, ID cards and letters never leave your device.
On clean printed Hindi it gets more than 99% of characters right, and a full A4 page takes under a second on most GPUs. For phone photos, keep the page flat and well lit. Boxes marked in amber were read with low confidence and are worth a quick check.
Not yet. It is trained on printed text: books, notices, bills, screenshots, forms. Handwritten notes usually come out garbled.
Yes. Drop a PDF and each page is rendered and read in turn (up to 30 pages at a time), with a thumbnail strip to check each one. Copy a page, copy everything, or download all the pages as one .txt file.
No limits, no signup, no watermark. It runs on your own device, so there is no per-page cost for us to charge you.