dstackai/dstack

dstack 是一个统一的 GPU 配置和编排控制平面，适用于任何 GPU 云、Kubernetes 或本地集群。

它简化了开发、培训和推理流程，并且兼容任何硬件、开源工具和框架。

加速器

dstack 开箱即用支持 NVIDIA、AMD、Google TPU 和 Tenstorrent 加速器。

最新消息 ✨

• [2026/08] dstack 0.21.0：Pydantic v2，网关副本
• [2026/07] dstack 0.20.29：预设 — 代理驱动推理优化（实验性）
• [2026/07] dstack 0.20.27：Slurm 后端
• [2026/05] dstack 0.20.21：Kubernetes 多集群
• [2026/05] dstack 0.20.20：NVIDIA Dynamo 集成
• [2026/04] dstack 0.20.17：Kubernetes 卷
• [2026/02] dstack 0.20.10：PD拆分支持
• [2026/01] dstack 0.20.7：副本组

它是怎么运作的？

启动服务器

在通过CLI或API使用dstack之前，先设置一个dstack服务器。如果你已经有一个正在运行的dstack服务器，只需安装CLI。

要在GPU云或Kubernetes集群间协调计算，你需要配置后端。

dstack is a unified control plane for GPU provisioning and orchestration that works with any GPU cloud, Kubernetes, or on-prem clusters.

It streamlines development, training, and inference, and is compatible with any hardware, open-source tools, and frameworks.

Accelerators

dstack supports NVIDIA, AMD, Google TPU, and Tenstorrent accelerators out of the box.

Latest news ✨

• [2026/08] dstack 0.21.0: Pydantic v2, Gateway replicas
• [2026/07] dstack 0.20.29: Presets — agent-driven inference optimization (experimental)
• [2026/07] dstack 0.20.27: Slurm backend
• [2026/05] dstack 0.20.21: Kubernetes multiple clusters
• [2026/05] dstack 0.20.20: NVIDIA Dynamo integration
• [2026/04] dstack 0.20.17: Kubernetes volumes
• [2026/02] dstack 0.20.10: PD disaggregation support
• [2026/01] dstack 0.20.7: Replica groups

How does it work?

Launch the server

Before using dstack through CLI or API, set up a dstack server. If you already have a running dstack server, you only need to install the CLI.

To orchestrate compute across GPU clouds or Kubernetes clusters, you need to configure backends.

When using dstack with on-prem servers, backend configuration isn’t required. Simply create SSH fleets once the server is up.