Quickstarts
Follow the quickstarts in this guide to train models, run inference, or stand up your own Crusoe infrastructure. First, select a tab below based on how you'd like to host your work:
- Managed AI: Run inference or train a model and have Crusoe manage the infrastructure for you.
- Infrastructure Cloud: Provision and operate your own compute on Crusoe.
- Managed AI
- Infrastructure Cloud
| Quickstart | Description |
|---|---|
| Serverless Inference | Generate an API key and send requests to hosted models over an OpenAI-compatible API |
| Self-Serve Deployments | Run inference on reserved GPU capacity that Crusoe manages end-to-end |
| Serverless Fine-Tuning | Upload a dataset, launch a LoRA fine-tuning job, and download the resulting checkpoint—no GPU infrastructure required |
| Quickstart | Description |
|---|---|
| Install the CLI | Install and authenticate the Crusoe CLI |
| Get started with Terraform | Manage infrastructure programmatically using Terraform |
| Create a VM | Provision a GPU or CPU VM and connect via SSH to run workloads |
| Spin up a GPU cluster | Provision a multi-node GPU cluster using Kubernetes or Slurm |