Ollama
Self-hosted LLM inference on a GPU node pool with models pulled at start.
About
Self-hosted LLM inference on a GPU node pool with models pulled at start.
This is a placeholder for a template we are considering. It is not in the library yet. If you want it, press Deploy and tell us; you can also deploy the project today from your own repository with the same plan, policy and approval flow.
Use cases
- Private model serving for internal tools.
- Evaluating open models without sending data out.
Inputs
Inputs will be defined with the template. Expect cluster or cloud account, environment name, sizing and a secret for the admin credentials.
What it creates
How it runs on StackGuardian
Runs as a stack: infrastructure first, then the application release, each step with its own plan, policy check and approval.
Approvals happen in StackGuardian. An agent can never approve its own run. Every run keeps its plan, policy results and approver in the audit log.