Neural network-driven streaming voice-over translation is a complicated technological task that nobody in the world has managed to achieve yet. Unlike recorded videos, where a neural network immediately is fed the full audio track, streaming requires work on the fly, like a simultaneous interpreter.

This year, we’ll be streaming YaTalks simultaneously in two languages using these technologies. We believe our solution will help users overcome at least some language barriers and discover a whole new world of content that hasn’t been translated yet.