I really like the idea here, and am all for more people using the canvas element, but I don't really think the built in canvas API is unwieldy, or even remotely foreign to anyone working on the web.
The only interesting functionality exposed in the API that I could find was some basic animation - defining a translation, saving the context of the canvas (for later rendering?) and then running the animation. Can anyone explain why this is interesting? Maybe it's a little early to pass judgement?
edit: Okay, the support for tags on objects you draw onto the canvas is pretty useful. This would definitely be useful for anyone developing a canvas game.