> ## Documentation Index
> Fetch the complete documentation index at: https://docs.compute.cx/llms.txt
> Use this file to discover all available pages before exploring further.

# Introduction

> One CLI. A fresh GPU machine per run. Prepaid credit, billed while the machine exists.

Compute runs a Python function on a newly provisioned cloud GPU, streams output to your terminal, returns JSON, and confirms the machine is gone.

You do not manage images, keep-alive pools, or a login to the box. The machine exists for one run.

## What you do

```bash theme={null}
curl -fsSL https://compute.cx/install.sh | sh
compute setup
compute credits add 10
compute run train.py::main --gpu H100-SXM --wait
```

`--gpu H100` is an alias for **H100-SXM**.

The three-command funnel (install → setup → run) also lives on [compute.cx/quickstart](https://compute.cx/quickstart). This site is the manual.

## What is live

| Surface                    | Status                                             |
| -------------------------- | -------------------------------------------------- |
| Install (`curl\|sh`)       | Live                                               |
| Sign-in + `compute setup`  | Live                                               |
| Prepaid credits            | Live — \$10 minimum                                |
| RunPod Secure **H100-SXM** | Public lead SKU                                    |
| Hot Aisle **MI300X**       | Available, stock can be tight                      |
| Cancel + teardown          | Live                                               |
| Dashboard                  | Observe keys, credits, runs — **not** the launcher |

Compute is in early access. One active run per account and spend or create-rate limits may refuse a run; provider capacity is not guaranteed. Different accounts are not serialized by a Compute-wide machine slot.

## What this manual does not cover

These exist in design or in the binary and are **not** documented as customer-ready here:

* `pip install compute` (v0.2)
* Detached SDK handles that you wait on later
* Persistent disks
* Checkpoint resume
* Community / interruptible / extra consumer SKUs
* Multi-GPU shapes

Rates and the platform fee are on [Pricing](https://compute.cx/pricing). Use `compute gpu list` for the locked list rate on a SKU. Do not hard-code a fee percentage from this page.

## Next

<CardGroup cols={2}>
  <Card title="Install" href="/get-started/install" icon="download">
    Isolated venv, `compute` on PATH.
  </Card>

  <Card title="First run" href="/get-started/first-run" icon="play">
    Tiny MLP on H100-SXM.
  </Card>
</CardGroup>