To me, it sounds like you have the perfect head start to this whole scene. I can see why you want people to know this exists...
Your quoting "live" there made me hesitate on the word for just long enough to start wondering about realtime processing, which I get the impression your system is capable of (if you were able to do SD card tests and instantly preview the results).
Now I'm wondering what would happen if you ran your software on a box in the cloud with a 1Gbit link, where people feed in their dual- or tri-camera raw footage, and perfect 3D (and maybe with a hefty premium, calculated depth info) comes out the other end - all in near-realtime. Can you manage less than 1s of latency? 500ms even? (Okay maybe I'm being too optimistic here)
Hmm, and maybe you could provide direct ingress support for broadcast packs....
And thinking about it, depending on how CPU-intensive your system is, you might be able to run this on relatively inexpensive hardware per 1G/2G link. (I know EPB offers commercial 10G installation in Chattanooga for $300/mo, but I'm not sure what their datacenter pricing is.)
I really like the idea of post-processing as a service - considering what options are on the table, that direction seems the least complex and the most straightforward.
(PS. Please excuse the rather non-provisional and overly concrete/direct/instructional wording style of my previous post - my brain's a little under the weather at the moment, and grammaring how I'd prefer to is proving occasionally challenging.)