Multi-user browser-based viewing in Omniverse

Hi everyone,

I’m exploring ways to operate and view an Omniverse environment remotely.

I’ve been experimenting with the usd_viewer sample from kit-app-template,
which works very well for one-to-one WebRTC streaming.

Now I’d like to expand this to the following scenario:

  • A presenter operates an Omniverse scene during a lecture or live event.

  • Audience members access the same scene via their web browsers, without installing Omniverse.

  • Each viewer can independently control their own camera view while watching the same scene.

Question:
What would be the best approach to realize this kind of multi-user, browser-based viewing environment in Omniverse?
Even if it’s not easily achievable right now, I’d also like to know what kinds of technical preparation, infrastructure, or NVIDIA services could make it possible in the future (e.g., multi-session streaming, collaborative scene sync, or server scaling).

Any advice, examples, or architectural pointers would be greatly appreciated.

Thanks in advance!

Whilst this is technically possible, you would need to mount your project file on an NVIDIA Cloud Server like GDN or something similar, and then for every person there, you would need an RTX instance running. That would require a lot of cloud compute and credits. Or it could be your own AWS or Azure Server. Either way, possible but quite a major investment. What would be better is you just share the file publicly and each user can download and open it on their own laptops. That is simple and easy, if they have modern PC laptops.

Graphics Delivery Network (GDN) for Interactive 3D Experiences | NVIDIA