I didn’t know about it; it looks interesting and I’ll give it a whirl when I can! 
For now, the main difference I see between the two would be that such functionality, integrated into the website would allow one to work on a mission and adjust settings on the fly, without having to use multiple platforms or intermediate steps.
Moreover, simulating the behaviour of the drone in terms of how settings are interpreted through Litchi’s algorithms would be the main benefit of the interface. To put it in my own example - I came to this idea from the perspective of an issue I have, where the drone turns abruptly at some point of the flight and I can’t tell why. In cases like this tweaking settings one by one is time and battery consuming, and in the end you don’t know if it isn’t just the drone not being able to follow a type of instruction the app is giving.