Additional Deployments

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

VSS can be deployed in several ways depending on how quickly you want to get started and how much control you need over the configuration. This page summarizes the available options, from the native Docker Compose quickstart to fully customizable Docker Compose and Helm Chart deployments.

Docker Compose Quickstart

The Quickstart deploys a basic vision agent with native Docker Compose. Each profile keeps deployment-specific settings in a user-selected *-overrides.env file (the examples use user-overrides.env); the other agent workflows build on the same Compose pattern for video summarization, search, and alerts.

Other Deployment Options

• Brev Launchable - A preconfigured, fully optimized NVIDIA Brev compute and software environment that uses the Docker Compose method to stand up the blueprint with minimal setup.
• Edge Deployment - Deploy the base vision agent on edge platforms (DGX Spark, IGX Thor, AGX Thor) using edge-optimized Nemotron LLMs.
• Helm Deployment - Deploy VSS developer profiles on Kubernetes using Helm, the fully customizable alternative to Docker Compose for production and multi-node clusters.
• Secure Deployment - A network-segmentation recipe to follow before exposing VSS in environments where untrusted users, services, or networks may be present.