I've just launched https://www.verbascan.com/ — it does context-aware photo translations. Still early (doesn’t ask questions yet), but even at this stage, it's clear how much better the results are when intent and nuance are factored in. Would love any feedback if you give it a try.