I think the dream that Macaw was selling was that a designer can "draw" code. So designers who are used to tools like Photoshop and Sketch can draw boxes on the Macaw canvas and it will output "semantic" code. When a tool shows off that they export clean, semantic code, they are essentially saying that it's production-quality (beyond just a mockup).
I saw a few tweets from Macaw saying that they're a prototyping tool, but that messaging didn't go beyond tweet replies. Their entire marketing message was about "drawing" beautiful, semantic html/css that's ready to be handed off to a developer or to push live on a server.
What is Scarlet? I think it's a tool for front-end coders and for designers that have experience with coding websites. I think they're giving up on the promise that code can be drawn, which was the #1 thing that got all types of multi-disciplinary designers excited and hopeful. The Scarlet landing page states "Was this site built in Scarlet? Indeed, it was! However, it was designed by someone with an understanding of HTML, CSS and JS. That is a prerequisite for making the most of Scarlet." That's to say that you don't just need to understand these languages, you have to be able to write them pretty proficiently to create similar web experiences. (Check out the JS file, which is written by an experienced programmer: http://scarlet.macaw.co/js/scarlet.js). I personally think Scarlet will appeal to some designers out there that like to code, but i'd argue that it's a smaller percentage of the people on their email list.
Webflow's mission is to empower designers/entrepreneurs/pseudo-coders to build the same stuff that experienced front-end/back-end developers build, but do it way faster and without touching a line of code. One of the biggest hurdles currently in Webflow is understanding web design basics like the box model and working with classes/selectors. But as soon as new users get those concepts they are empowered to design awesome stuff. They don't have to write code. I think theres a big difference there. Down the road we'll have some type of code integration that makes sense, but at the moment we're trying to solve a harder problem. (Visual coding? haha)