A lot of real-world documents in India aren't purely in English — a signboard, a handwritten note, or a government form often mixes English with Hindi or a regional script in the same image. Most free OCR tools online only recognize English by default, which means that mixed text either gets garbled or silently dropped. Our Image to Text (OCR) tool now has a language picker to handle this properly.
What's supported
Alongside English, you can now recognize Hindi, Bengali, Tamil, Telugu, Marathi, Gujarati, Kannada, Malayalam, Punjabi, Urdu, Odia, and Assamese — each paired with English by default, since that's how most real documents actually mix scripts, rather than being purely single-language.
Why this is still entirely in your browser
Each language is its own separate recognition model, so rather than downloading all thirteen up front, the tool only fetches the one you actually select, the first time you use it in that language — after that, your browser caches it, and later runs in that same language are fast. Nothing about this changes the tool's core privacy property: the image and the recognized text never leave your device.
Where this matters most
The same language picker is also built into the new Indian Document OCR tool, since Aadhaar cards, PAN cards, and driving licences frequently include a regional-language line above or alongside the English text.
Using the tool
Pick a language before you run the tool — choosing an Indian language downloads that language's recognition data on first use, cached after that. Try it here: Image to Text (OCR).