NVIDIA
NVIDIA
nsight-copilot
Container
NVIDIA
NVIDIA
nsight-copilot

NVIDIA Nsight Copilot backend for offline self-hosted deployment on DGX Spark

In offline mode the server runs entirely on a local DGX Spark, routing all inference requests through an on-device LiteLLM proxy without sending prompts or code to any external service.

Publisher
NVIDIA
NVIDIA
Latest Taglatest
UpdatedJuly 19, 2026 UTC
Compressed Size156.33 MB
Multinode SupportNo
Multi-Arch SupportYes