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

# AI inference playground

> Test LayerRail-hosted chat and embedding models from the console, compare outputs, and copy generated code snippets to use in your app.

The inference playground lets you send test prompts to an available text generation endpoint.

## Start a chat

<Steps>
  <Step title="Open AI inference">
    In your project, open **AI Inference**.
  </Step>

  <Step title="Create an API key">
    Create an inference API key if the project does not already have one.
  </Step>

  <Step title="Open Playground">
    Choose the **Playground** tab.
  </Step>

  <Step title="Pick a model and key">
    Choose an inference endpoint and an API key.
  </Step>

  <Step title="Send a message">
    Enter a prompt and submit the request.
  </Step>
</Steps>

## Advanced settings

Advanced settings can include:

* Temperature.
* Top-p.
* Maximum tokens.
* Streaming mode.
* System prompt.

Use lower temperature for deterministic answers and higher temperature for more varied responses.

## Files

If file upload is available for the selected endpoint, attach files only when the model supports that input type.

<Info>
  Embedding endpoints are not chat models. Use text generation endpoints in the playground.
</Info>

## Troubleshooting

| Error           | What to check                                          |
| --------------- | ------------------------------------------------------ |
| `400` response  | Confirm the selected model supports the request shape. |
| Missing API key | Create an inference API key first.                     |
| Empty response  | Try a shorter prompt or a different model.             |
| Unauthorized    | Make sure the key belongs to the current project.      |