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

# Host

> Understand the costs and responsibilities of running an open-weight model.

Run an open-weight model on a cloud GPU in your own account. **Iwant** provisions the server and returns an OpenAI-compatible endpoint. You pay for cloud resources rather than buying inference at a provider's token rate.

## Launch your first model

You need `gcloud` and a GCP project with billing enabled, L4 GPU quota, and available capacity. Install Iwant and authenticate:

```bash theme={null}
curl -fsSL https://github.com/narevai/iwant/releases/latest/download/install.sh -o /tmp/iwant-install.sh
bash /tmp/iwant-install.sh

gcloud config set project YOUR_PROJECT_ID
gcloud auth login
gcloud auth application-default login
iwant auth
```

The installer puts `iwant` in `~/.local/bin`; add that directory to your `PATH` if needed. Preview the deployment, then launch:

```bash theme={null}
iwant up gpt-oss-20b --infra gcp --dry-run --yes
iwant up gpt-oss-20b --infra gcp --yes
```

Save the **Server** URL and **API key** printed after the model is ready. Iwant is free; GCP charges for provisioned resources, including idle GPU time.

Remove the deployment when you finish:

```bash theme={null}
iwant down
iwant list
```

Confirm the cluster no longer appears. Ctrl+C does not stop a remote server.

## Learn more

[Explore Iwant](https://www.narev.ai/docs/iwant/index.md) and use the [quickstart](https://www.narev.ai/docs/iwant/quickstart.md) to send your first inference request. See [launch options and recipes](https://www.narev.ai/docs/iwant/usage.md) for deployment settings.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.