Regarding key bindings, I'm planning on adding some soon, so that it can be fully interacted via keyboard (not sure if I'll use vim bindings exclusively — non-geeks don't know about vim).
On pandoc integration, it's complicated, since the server-side is implemented in node.js. Anyway, my To-Do has an item on supporting markdown :)
Finally, this won't be commercialized. As long as I have few-to-none costs supporting the hosting, it'll be free forever (i.e., why the heck people have to pay for these kind of editors?)
Indeed, the chosen design is a compromise between simplicity and practicality (I just couldn't imagine having a file/document browser taking up much space). Nevertheless, I'm thinking on having 0-9 shortcut keys to select documents.
Regarding dimming the word/char count, you're right. Will be fixed on the next iteration!
As it's been said, it's exactly the point. The current keyboard mappings are required to trigger fullscreen, while (non-alpha & non-beta) browsers don't implement a fullscreen API.
* mouse-clickable button: APIs for handling the fullscreen state of browsers are still nascent (Chrome 15 and Firefox 9-ish?), so it's a matter of months until adding it;
* Chrome Web store: already on the to-do (plus 100% offline usage via HTML5 manifest).