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

# HOSTED INFERENCE

> Deploying your trained model as a live API endpoint

## OVERVIEW

Once you’ve [generated training data](/guides/datasets/create), [fine-tuned a model](/guides/training), and [evaluated its performance](/guides/evaluations), the final step is deployment. Hosted inference lets you take a model trained on the Oumi Platform and serve it as a live API endpoint, making it available for real-time use.

<Note>The Deployments feature is currently in beta.</Note>

Each deployed model is assigned its own dedicated endpoint for inference, and you retain full control over its lifecycle, with the ability to create or remove deployments as needed.

## ACCESSING DEPLOYMENTS

<Info>To deploy a model, you’ll first need a trained model in your project to enable hosted inference.</Info>

* From the top of the **Models** page, click on the `Deploy Model` button; alternatively, click on the `+ Create Deployment` button from the **Deployments** page.
* On the **Deploy Model** modal window, select either `Custom Oumi Model` or `External Model`:

### Custom Oumi Model

* Provide a unique `Deployment Name`.
* Select a `Model` from the drop-down.
* Click `Start →` to deploy your model.

### External Model

* Provide a unique `Deployment Name`.
* Select a `Provider` from the drop-down.
* Select your `External Model` from the drop-down.
* Insert the API key for your provider.
* Click `Start →` to deploy your model.