Get Started With NVIDIA Run:ai

Choose how to build with NVIDIA Run:ai open source projects, deploy the full platform from the NVIDIA NGC™ catalog, or connect with NVIDIA for guidance and support.

Build With Open Source

Explore NVIDIA Run:ai projects on GitHub—including KAI Scheduler, KAI GPU Sharing Operator, and NVIDIA Run:ai Model Streamer—and use them independently or together in your Kubernetes platform.

Visit Github

Get the Full Platform From NGC

Download and deploy the complete NVIDIA Run:ai platform from NGC.

Get NVIDIA Run:ai From NGC

Contact Us

Talk with NVIDIA about supported deployment and procurement options, including NVIDIA DSX™ OS, NVIDIA Mission Control™, NVIDIA AI Enterprise, standalone, cloud marketplaces, and partners.

Contact NVIDIA

FAQs

With NVIDIA Run:ai 3.0, the complete NVIDIA Run:ai platform is available as open source software for AI workload and GPU orchestration on Kubernetes.

No license is required to deploy and self-operate the open source software. Licensing or entitlement terms apply when you choose an NVIDIA-supported commercial offering.

No. NVIDIA-supported options provide a commercial support and procurement path for the open source platform rather than a separate proprietary feature tier.

Choose an eligible NVIDIA-supported path:

  • NVIDIA Run:ai standalone
  • NVIDIA AI Enterprise
  • NVIDIA Mission Control
  • Participating cloud marketplaces and NVIDIA Run:ai partners

Yes. Builders can choose the level of adoption that fits their platform:

  • Deploy the complete NVIDIA Run:ai platform.
  • Use an individual project or combine projects with an existing Kubernetes platform.
     
Products
Technologies
Resources
Company Info