Compute with Hivenet lets you run on-demand GPU and CPU machines for AI inference, training, and general workloads. You create an instance, choose the resources you need, and connect over the network just like any other server.
Heads up: Compute uses a separate account from the main Hivenet app. If you already use Store, you’ll still need to sign up for Compute.
What you can do
Spin up on-demand GPU instances for AI workloads
Host inference servers (for example, vLLM)
Use templates to start with preinstalled runtimes and frameworks
Connect via SSH or web console to manage your instance
What you’ll need
A Compute account
A payment method added in Billing
An SSH key or your preferred connection method
Quick start
Sign in to Compute with Hivenet.
Select Create instance.
Pick a template or start from a clean image.
Choose resources, then create the instance.
Connect and run your workload.
Tip: Start small. You can create a second, larger instance later if your job needs more resources.
Warning: Billing accrues while an instance is running. Terminate the instance when you’re done to stop charges.
Pricing and billing basics
You’re billed for active runtime per instance.
Adding a payment method is required before creating paid instances.
Terminating an instance stops future runtime charges.
Data and security
Compute runs workloads on Hivenet’s distributed infrastructure.
You control what runs on your instance and who can access it.
For details on network access, data handling, and compliance, see the security section in our docs.
Learn more in the docs
The full step-by-step guides live in our docs. You’ll find:
Quickstart and setup
Creating and connecting to an instance
Inference servers (vLLM)
Templates and images
Billing and invoices
Security and access