What’s the distinction you’re drawing here? I don’t know the implementation details, but my understanding is that it’s fundamentally exactly the same thing: the map is rendered by the browser/client using lines and polygons, rather than loading pre-rendered tiled images.