> This page is for version Dev.
> For other versions, use one of these documentation indexes:
> - 3.3.0 (default): https://docs.nvidia.com/vss/3.3.0/llms.txt
> - Dev: https://docs.nvidia.com/vss/dev/llms.txt

> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/vss/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/vss/_mcp/server.

# Known Limitations

* Only the default NIMs (LLM and VLM) have been tested to run on shared GPU. Use Dedicated GPUs or Remote NIMs for other models.

* **WebRTC based video playback in the UI**: In the reference Video Analytics UI, video playback may not be available. If the video streaming service runs behind a private network where the WebRTC TURN server is unreachable or not set up appropriately, playback fails while the rest of the pipeline continues to work normally.

* VSS UI queries targeting multiple sensors are expected to take several tens of seconds.

* The Cosmos3 Reasoner NIM and VIOS microservices may crash over long duration runs.

* In the event that the [RTVI CV](/system-components/vision-microservices/object-detection-and-tracking) or VIOS microservice crashes/restarts, streams may no longer be processed. To get around this, either manually restart the `perception-sdr-smc` container or delete and re-add all streams in the VST UI.