1. Home

NVIDIA SDK Manager

NVIDIA SDK Manager provides an end-to-end development environment setup solution for NVIDIA’s Jetson, RAPIDS, RTX Kit, Holoscan, DeepStream, Rivermax, GXF Runtime, Video Codec, Aerial Research Cloud (ARC-OTA), Ethernet Switch SDKs for both host and target devices.

Download Now
Documentation

How SDK Manager Works

NVIDIA SDK Manager development environment

NVIDIA SDK Manager automates development environment setup by downloading SDKs, BSPs (Board Support Packages), drivers, tools, and sample code directly from NVIDIA repositories.
For the host machine, it installs required packages including SDK components like CUDA, TensorRT, and Nsight tools, and resolves dependencies to ensure compatibility.
For embedded target devices, it flashes the OS (such as Linux for Jetson), installs drivers and SDK components, and manages dependencies over USB or network connections.
Supporting both GUI and command line modes, NVIDIA SDK Manager streamlines the entire setup process to deliver consistent, validated environments for AI, edge, and accelerated computing, reducing manual effort and setup time.

Key Features

img-alt-text

Centralized Hub for NVIDIA SDKs and Tools

Gain streamlined access to NVIDIA's extensive software ecosystem from a single point of control. SDK Manager serves as a unified hub for discovering, installing, and managing multiple software development kits, ensuring you have the needed versions at your fingertips.

img-alt-text

Streamlined End-to-End Development Environment Setup

SDK Manager automates the entire workflow from a bare host machine to a ready-to-use development environment. It orchestrates the entire setup in one seamless workflow, handling tool chains, SDKs, and essential system configuration, minimizing manual setup efforts and reducing the risk of errors.

img-alt-text

Intelligent Dependency and Compatibility Management

SDK Manager automatically resolves the complex software dependencies between SDKs, OS drivers, libraries, and tools. It ensures all selected components are cross-compatible and work together seamlessly, preventing version conflicts and creating a stable, reproducible environment for development and testing.

img-alt-text

Unified Management for Host and Target Environments

Manage your entire cross-platform development pipeline from a single application. SDK Manager seamlessly handles installation on your host machine and flashing for a wide range of NVIDIA target hardware, supporting a wide variety of operating systems on both host and target devices. This enables efficient development and deployment in a broad range of platforms and configurations.

img-alt-text

Flexible Interface for Optimized Workflows

Choose the interface that best suits your development style. SDK Manager offers an intuitive Graphical User Interface (GUI) for straightforward setup and a powerful Command-Line Interface (CLI) for automation and scripting, supporting a variety of development flows.

img-alt-text

Proactive Update and Maintenance Notification

Stay current with the latest NVIDIA software advancements. The tool provides automatic notifications for available updates to SDKs, drivers, and tools, ensuring your development environment benefits from the latest features, security patches, and performance improvements.

Host Operating System Compatibility Matrix

While SDK Manager supports all the below host operating systems, you need to verify the SDK package supports the host OS; these requirements are set by the SDK package you are installing.

Base SDKs Host Operating System Compatibility Matrix

See the Base SDKs host operating system compatibility matrix for details:

NVIDIA SDK SDK Version Supported Host Operating System
Ubuntu CentOS Windows
16.04 18.04 20.04 22.04 22.04 (ARM) 24.04 8.2 10 11
JetPack JetPack 4.x ✔ ✔
JetPack 5.x ✔ ✔
JetPack 6.0.x ✔ ✔
JetPack 6.1.x or 6.2.x ✔ ✔ ✔ 1 ✔ 2 ✔ 2
JetPack 7.x ✔ ✔
Ethernet Switch Switch 4.x ✔ ✔ ✔
ARC-OTA Aerial 22-4 ✔
Aerial 23-4 ✔ ✔ ✔
ARC-OTA 1.x ✔ ✔ ✔ ✔ 3
RAPIDS RAPIDS 24.x ✔ ✔ 4
RAPIDS 25.x ✔ ✔ ✔ 5
Holoscan HoloPack 1.2 ✔
RTX Kit RTX Kit 2025.4 ✔ ✔

¹ Support Jetson device flashing and target components installation only. Host development is not supported.
² Supported from Jetpack 6.2.1
³ Supported from ARC-OTA 1.6
⁴ Supported from RAPIDS 24.12
⁵ Supported from RAPIDS 25.04
* For older SDK versions run: sdkmanager –-archived-versions
* DRIVE OS: The latest version of SDK Manager no longer supports DRIVE OS. To install DRIVE OS 5.0.x or 6.0.x, use SDK Manager 2.2.0 from the archive page.
* DOCA: The latest version of SDK Manager no longer supports DOCA. To install DOCA 1.5 or 2.2, use SDK Manager 2.1.0 from the archive page.

Additional SDKs - Supported Base SDK Compatibility Matrix

See the supported Base SDKs for each additional SDK:

NVIDIA SDK SDK Version Supported Base SDK
JetPack
4.x 5.x 6.x 7.x
DeepStream DeepStream 6.x ✔ ✔
DeepStream 7.x ✔
DeepStream 8.x ✔
DeepStream 9.x ✔
GXF GXF Runtime 4.x ✔ 1
GXF Runtime 5.x ✔
Holoscan Holoscan 2.x ✔
Holoscan 3.x ✔ ✔
Holoscan 4.x ✔ ✔
Video Codec Video Codec 13.0 ✔ 2
Rivermax Rivermax 1.31 ✔
Rivermax 1.4x ✔

¹ Available only on Ubuntu 22.04
² Supported from Jetpack 7.1

    What’s New in SDK Manager - Version 2.4.0

    • Support flashing and target installation from Windows hosts for JetPack 6.2.x and JetPack 7.x SDKs.
    • SDK Manager can now run directly on NVIDIA GH200 servers to install ARC-OTA SDK for ARM
    • Support RTX Kit installation on Windows host. Starting with RTX Kit 2025.4
    • Display release messages and update dialogs for additional SDKs, such as DeepStream and Holoscan
    • Resolved a performance issue that caused lag when scrolling through the embedded terminal with extensive output
    • Other bug fixes and stability improvements

      Get Started With NVIDIA SDK Manager

      The SDK Manager empowers developers to work seamlessly across platforms—whether you're coding on Linux, Docker, or Windows.

      DOWNLOAD and INSTALL

      Local
      • Download the install file for your desired platform
      • Run the command for your desired platform:

      Ubuntu (x86_64):

      sudo apt install ./sdkmanager_[version]-[build#]_amd64.deb

      Ubuntu (arm64):

      sudo apt install ./sdkmanager_[version]-[build#]_arm64.deb

      CentOS:

      sudo dnf install ./sdkmanager_[version]-[build#].x86_64.rpm
      Network

      From a terminal window, run the following commands for your desired platform:

      Ubuntu (x86_64):

      wget https://developer.download.nvidia.com/compute/cuda/repos/[distro]/x86_64/cuda-keyring_1.1-1_all.deb
sudo dpkg -i cuda-keyring_1.1-1_all.deb
sudo apt-get update
sudo apt-get -y install sdkmanager

# Replace the distro with the one that you are using. Supported distros are: ubuntu2404, ubuntu2204, ubuntu2004, ubuntu1804

      CentOS (x86_64):

      sudo dnf config-manager --add-repo https://developer.download.nvidia.com/compute/cuda/repos/rhel8/x86_64/cuda-rhel8.repo
sudo dnf clean all
sudo dnf install sdkmanager

      LAUNCH

      From a terminal window, launch SDK Manager with the command:

      sdkmanager

      DOWNLOAD and INSTALL

      • Download the installer that is compatible with Windows 10 and 11.
      • Run the installer to install SDK Manager application step by step.

      LAUNCH

      Graphical Interface (GUI):

      • Use the Start menu shortcut "NVIDIA SDK Manager" or
      • Run NvSDKManagerGUI.exe from the installation directory

      Command-Line Interface (CLI):

      • From a terminal window run:
      NvSDKManager.exe --cli
      • If not added to the path during installation: from a terminal window, go to the installation directory and run:
        • ./NvSDKManager.exe --cli

      For more information, see SDK Manager user guide (Download and Run section).

      DOWNLOAD and INSTALL

      • Download Docker image with the base OS that you need for required SDK
      • Install the Docker image:
        From a terminal, load the Docker image:
      docker load -i ./sdkmanager-[version].[build#]-[base_OS]_docker.tar.gz

      It is recommended that you tag the version as "latest" for ease of use:

      docker tag sdkmanager:[version].[build#] sdkmanager:latest

      LAUNCH

      • The Docker image is designed to be executed directly from the host, without the need to open the terminal inside the docker itself. The sdkmanager executable is the entrypoint.
      • SDK Manager CLI arguments should be used directly when running a new container. Example command: SDK Manager CLI with Docker:
      docker run -it --rm sdkmanager --help

      For more information, see SDK Manager user guide (Docker section).

      System Requirements

      System requirements of SDK Manager may differ from the system requirements of the specific SDK you are installing.
      See System Requirements for each SDK.

      Archives

      Resources

      SDK Resources

      Documentation

      NVIDIA Jetson

      NVIDIA Product

      NVIDIA Jetson™ is a powerful platform for developing innovative edge AI and robotics solutions across industries. It delivers the compact, energy-efficient modules and developer kits with a robust AI software stack you need to deploy next-generation physical AI solutions.

      Documentation

      NVIDIA Holoscan SDK

      NVIDIA Product

      NVIDIA Holoscan is a domain-agnostic, multimodal AI sensor processing platform that provides the accelerated, full-stack infrastructure needed for real-time processing of streaming data at the edge or in the cloud.

      Documentation

      NVIDIA Rivermax SDK

      NVIDIA Product

      NVIDIA® Rivermax® offers a unique IP-based solution for any media and data streaming use case.

      Documentation

      NVIDIA DeepStream SDK

      NVIDIA Product

      NVIDIA DeepStream’s multi-platform support gives you a faster, easier way to develop and deploy real-time video streaming pipelines for generative AI agents and applications. You can even deploy them on premises, at the edge, and in the cloud with just the click of a button.

      Documentation

      NVIDIA GXF Runtime SDK

      NVIDIA Product

      Get step-by-step instructions for deploying pretrained models as services on a local workstation and how to interact with them through a client.

      Documentation

      NVIDIA Video Codec SDK

      NVIDIA Product

      A comprehensive set of APIs including high-performance tools, samples and documentation for hardware-accelerated video encode and decode on Windows and Linux.

      Documentation

      NVIDIA Ethernet Switch SDK

      NVIDIA Product

      The NVIDIA® Ethernet Switch SDK provides the flexibility to implement any switching and routing functionality, with sophisticated programmability that doesn’t compromise performance in packet rate, bandwidth, or latency.

      Documentation

      NVIDIA RAPIDS SDK

      NVIDIA Product

      CUDA-X™ Data Science is a collection of open-source libraries that accelerate popular data science libraries and platforms. It is part of the CUDA-X collection of highly optimized, domain-specific libraries built on CUDA®.

      Documentation

      NVIDIA Aerial Resource Center (ARC-OTA)

      NVIDIA Product

      NVIDIA AI Aerial™ is a suite of accelerated computing platforms, software libraries, and tools to build, train, simulate, and deploy AI-native wireless networks. It enables developers and researchers to go from rapid prototyping to commercial development of AI-RAN solutions for 5G and 6G that telcos can deploy.

      Documentation

      NVIDIA RTX Kit

      NVIDIA Product

      NVIDIA RTX™ Kit is a suite of neural rendering technologies to ray trace games with AI, render scenes with immense geometry, and create game characters with photo-realistic visuals.

      Open Source Resources

      SDK Manager is using open source software available for download here.

      Everything You Need to Set Up Your Development Environment

      Get Started