Open-source, cloud-agnostic control plane for orchestrating AI workloads anywhere.
Visit Dstackdstack is an open-source, vendor-agnostic control plane that unifies fleets, dev environments, tasks, services, and volumes for containerized AI workloads. Built with a simple CLI, UI, and API, dstack orchestrates AI/ML workloads across any cloud, on-prem clusters, or marketplace GPUs, providing streamlined accelerator provisioning, workload scheduling, and infrastructure observability for both development and production inference. It is ideal for AI and ML teams seeking to unify and optimize compute resources across heterogeneous environments without the headaches of traditional orchestration tools.
Visit Dstack's official website for product details and getting started.
Comprehensive guidelines and API reference for using dstack.
Insights and updates on AI/ML workloads and dstack features.