OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Infrastructure & Serving

Dstack

Dstack is an open-source control plane for provisioning and orchestrating GPU compute across GPU clouds, Kubernetes, virtual machines and on-premises clusters. It provides configurations for fleets, development environments, training and batch tasks, inference services and persistent volumes, with workload settings defined in repository YAML and applied through a CLI, API or AI-agent skills.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Dstack — OpenViST project profile in AI Infrastructure & Serving.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Dstack when AI development, training or inference workloads need GPU compute provisioned and coordinated across cloud, Kubernetes, virtual-machine or on-premises environments.

Who it is for

  • AI developers who need managed GPU development environments.
  • Teams running training or batch jobs across GPU clouds, Kubernetes or on-premises compute.
  • Operators deploying model inference services on GPU infrastructure.
  • Teams managing GPU resources across more than one infrastructure environment.

Documented uses

  • Provision and manage compute fleets across supported infrastructure.
  • Launch development environments for IDE or agent access.
  • Run training, batch and other jobs on a node or across clusters.
  • Deploy model inference as endpoints.
  • Manage instance and network volumes for persistent data.
Before you choose

Requirements & limits

Dstack is available as a public-source repository under the Mozilla Public License 2.0.

  • A Dstack server must be set up before using the CLI or API.
  • GPU cloud and Kubernetes use requires backend configuration; on-premises servers can be managed as SSH fleets without backend configuration.
  • The documented server and CLI prerequisites include Git and OpenSSH.
  • The project lists NVIDIA, AMD, Google TPU and Tenstorrent accelerator support.
  • Presets for agent-driven inference optimization are identified as experimental.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up a local Dstack server and connect the CLI

Install the Dstack server and configure the CLI to address it.

Before you start

  • Git and OpenSSH
  • Linux, macOS or Windows with WSL 2
  • The documented server installation uses uv

Steps

  1. Install Dstack with the command below.

    Documented command
    uv tool install "dstack[all]" -U

  2. Start the server with dstack server.

  3. Install the CLI separately with the command below if it was not installed with the server.

    Documented command
    uv tool install dstack -U

  4. Configure the CLI with the server address, user token and project name using dstack project add.

Expected outcome

A running Dstack server and a CLI project configuration pointing to that server.

Guide limitations

  • This covers server and CLI setup only; it does not configure a compute backend or run a workload.
  • The server address and token must correspond to the server being configured.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for Model inference and serving

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
DstackProvision and manage compute fleets across supported infrastructure.Cloud, Docker, Self-hosted, ServerDstack is available as a public-source repository under the Mozilla Public License 2.0.
MorphingdbImport a libtorch model and create predictions from values in a PostgreSQL table.Docker, Linux, Self-hosted, ServerThe README documents a Docker image and a build-from-repository option, followed by connecting to PostgreSQL and enabling the pgdl extension.
Gateway — pyadavSend chat-completion requests to a provider through a common server endpoint.ServerThe repository is public and identifies the project as Apache-2.0 licensed. README installation instructions include binaries, platform package managers and Docker.
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
13 October 2026
Original website title
dstackai/dstack · Official homepage title · retained in its source language
How we review sourcesSuggest a correction