I wonder if the browser can do continuous recording to disk, I know Mega (mega.nz) does some sort of magic with client-side JS so it's not the browser just writing bytes it's getting from the HTTP connection, it's writing bytes to disk given to it by the JS. If the browser doesn't know how big the HTTP download is, it just writes and writes to disk, I wonder if that's possible with the fancy HTML5 save-to-disk.
This is something we are also investigating. Our goal is to have recording happen on the device, and save directly to the device. We have made some great breakthroughs with this but are still in the R&D phase. This is our ideal solution, though.
