We are using a cutting edge algorithm so please note that this first version is far from perfect. It is designed for handwriting and while it does well with clean text and most printed text, the result for extracting text in cursive writing is not great at this time and this currently only supports English. The system is also using a machine learning algorithm, that can make sense of objects in images and will provide some contextual information about what it sees.
If you are using the iOS app, please see our article here.
Note: This is currently in beta, and only supports English.