It would be interesting to experiment with putting a recording child process in the cloud as well as the replaying children. User interactions (e.g. navigations and mouse/keyboard events) would be sent to the recording process and the rendered data sent back to the user. The main advantage of this is that users on any platform could use web replay --- we would just have a recording process running on a mac in the cloud. The disadvantages are that pages hosted locally (offline) could not be debugged (as the recording process could not load them), that the look and feel of the page would be the same as if a mac were being used, and that the browsing experience would be laggier than having a local recording child.
I'm not sure whether this is a good idea, but noting it here as a potential design.