I will eventually write a blog post about it. But in the mean time this tweet give the overall gist

https://twitter.com/mattegreer/status/1244012323976515584

I made it so I can create a game scene out of ASCII characters, and then I can also tweak the scene and/or entities within the scene with code if needed (but it almost never is). Then the Storybook renderer instantiates a canvas and the game engine, and populates it with the scene.

Making your own renderer is pretty much totally undocumented. I took the React renderer as a basis (https://github.com/storybookjs/storybook/tree/next/app/react) and tweaked from there. I have spoken with Michael Shilman about this and there are plans to document it and make it easier to do, not sure where they are on that though.

It's so undocumented and uncharted I'm not even sure if "renderer" is the right name for these things.