Yeah, I use it as well and the level of translation is simmilar to the level of google translate in the early 2010s. The main problem it has is that the OCR is not great. It can only detect very clear horizontal texts and has problems in any other use case. Also it gives you a translation but no alternatives that might be good depending on the context but I think that might be outside the scope of this model.