A unified orchestration layer for heterogeneous AI compute. It standardizes how to manage compute and run training and inference on GPU clouds, Kubernetes, VMs, or bare-metal clusters.
dstack is a control plane that provisions and coordinates GPU computing for machine-learning development, training, inference, and agent workloads across cloud providers, Kubernetes, and on-premises clusters. It is used by teams running workloads on NVIDIA, AMD, Google TPU, or Tenstorrent hardware. The catalogue add-ons operate dstack through its CLI, API, or server workflow.
Latest release 0.22.5 · 22 Oct 2026
1 file for Codex and OpenCode: dstack AGENTS.md — 1,263 tokens loaded in every session.
AGENTS.md A 1,263 tok These files are dstackai/dstack's own configuration — they tell Codex and OpenCode how to work on this repository, so they are not mods to install elsewhere. Copy one as a starting point and replace the parts that are about this project.