Can I run a virtual machine, VPS or GPU server on Catalyst, with root access and a static IP address?

No. Catalyst is a serverless platform. It doesn’t offer virtual machines, VPS or dedicated servers, root or SSH access, GPU instances or a static IP address for your project. The nearest fit for software you would run on your own server is Catalyst AppSail, a fully managed platform for web services.

With AppSail you deploy your app as a Java, Node.js or Python build, or as a container image (an OCI image built for Linux AMD64) for any other runtime, such as PHP or Go. Catalyst spawns and scales the instances. You set the memory, disk size, port, startup command and environment variables, but you don’t manage the infrastructure. Check AppSail’s limits first: each request must complete within 30 seconds, and inactive instances are scaled down, so a process that must run around the clock, such as a blockchain node or a long-running worker, doesn’t fit.

For relational data, use Catalyst CloudScale Data Store instead of your own database server. For AI features, Catalyst QuickML LLM Serving gives you large language models served by Catalyst; Catalyst doesn’t provide GPU capacity to train or host your own model.

Refer to the AppSail Basics help page for details.