A menu in a language you don't read, a sign in a foreign city, or a document written in another language are all situations where a photo is easy to take but the meaning stays locked away. This tool combines two steps into one: it first reads the text visible in an uploaded image using optical character recognition, and then translates that recognised text into a language you choose, so you go from a photo straight to a translation without any manual typing in between.
To use it, upload or drop a photo containing text, choose the language you want it translated into from the dropdown menu, and press Read & Translate. The tool first scans the image and recognises whatever text it can find, shows you that original recognised text so you can sanity-check it against the photo, and then sends that text off to be translated, displaying the translated result underneath, ready to read or copy.
It's worth understanding the two-step nature of this honestly: the accuracy of the final translation depends heavily on how well the original text was recognised in the first place. Clear, well-lit photos of printed text in a common font produce reliable recognition and, in turn, a reliable translation; blurry photos, unusual handwriting, or highly stylised fonts can cause the recognition step to misread a word here and there, which then carries through into the translation. Straightening the photo, cropping tightly to just the text, and photographing in good lighting all make a real difference to the final result.
This is useful for reading a foreign-language menu or sign while travelling, getting the gist of a document or letter written in a language you don't speak, checking the meaning of text on a product's packaging, or translating a screenshot of a message or webpage that was not in your language to begin with. It's designed for a quick, practical translation of everyday text rather than for anything where legal or medical precision is required — for important documents, a professional human translator is always the safer choice.
For the most reliable results, try to photograph text that fills a good portion of the frame rather than a tiny amount of text lost in a large, cluttered photo, since giving the recognition step a bigger, clearer target to work with directly improves the quality of everything that follows, including the final translation.
The text recognition step runs entirely inside your browser using JavaScript on your own device. The translation step itself does need to reach a translation service over the internet to return a result in another language, since translation genuinely requires a language model or dictionary far too large to run offline — but the image itself is never uploaded anywhere, only the recognised text is sent for translation. The tool is free to use, with no sign-up and no limit on how many images you translate.