What it does
GPUStack is software for managing clusters of graphics cards (GPUs), the chips that run AI models. It can manage machines in your own building, in Kubernetes setups and with cloud providers.
It sets up and tunes engines such as vLLM, SGLang and TensorRT-LLM to serve text, voice, image and video models. Its makers say new models can be deployed on the day they are released. It can also start GPU machines that developers log into directly.
It includes user accounts, access control, monitoring and tracking of usage. A paid enterprise edition adds extra views.
Who it's for
IT teams and companies that run AI models on their own GPU servers for many users.
Can I use it without coding?
Developers only. It is server software installed with Docker on Linux machines with graphics cards. New to this? Start here.
Can I run it?
- Cost
- Free, with optional paid extrasThe open-source version is free; the makers also sell a hosted or pro version.
- Runs on
- Linux, Server, Cloud
- Computer
- Made for servers or data centres
- Privacy
- Runs on your own device: your data stays with you
- API key
- Not needed
Try it
- Prepare a Linux machine with at least one NVIDIA graphics card, its driver, Docker and NVIDIA's container toolkit.
- Start the GPUStack server with Docker as shown in the README.
- Use its web interface to add worker machines and deploy models.
Using Claude Code, Cursor or another AI coding assistant? Paste the prompt and it will walk you through the setup.New to this?
Before you install
- Only download from the official GitHub page. Copies with similar names exist for many popular projects.
- License (Apache-2.0): Permissive: free to use, change and share, including for business.
- More on staying safe