# Use AI Audio with your AI

> Connect Claude, ChatGPT or Grok to AI Audio through Geenius OS: it queues sound runs, reads takes and manages collections. A page makes them.

Source: https://aiaudio.geenius.io/docs/use-with-your-ai/

AI Audio has no AI you must use, and today its sound comes from a simulated Preview. Your own AI, Claude, ChatGPT or Grok, can queue sound runs, read the takes, and keep collections through Geenius OS. **A page makes the sound**: a run is made when somebody has AI Audio open.

> [!HONEST]
> The Preview provider is simulated in a browser: a robot voice, synthesised music beds and effects. It is not a voice model. Real voice models are connected in code and ready only where their keys are set. Making sound is on Premium and Max; Pro is the library. The door is on every plan, and during early access people request access.

## Connect it

Add AI Audio to your Geenius OS workspace and connect your AI to Geenius OS once, by the steps in the Geenius OS docs. Your AI then has tools called `audio` and `collection` of `geenius_geeniusaudio`.

## How a call looks

```json title="See what can make sound here"
{ "tool": "audio", "action": "providers", "args": {} }
```

## Your first ask

```text title="A first ask"
Use Geenius OS and AI Audio. Check which providers are ready. Then queue one music bed, calm, 30 seconds, as a loop, with the Preview, tell me what it costs and tell me to open AI Audio so it gets made.
```

## The tools

::mcp-tools

Ask for the actions with `action: "help"`, and for one action with `of`.

## Habits that work

- **Ask for providers first.** `providers` says what is ready.
- **Read the cost.** `create` answers the credits a run costs.
- **Queue, then say who makes it.** Tell the person to open the app.

## What your AI cannot do

It cannot make sound without an open page for the browser Preview, and it cannot make it on Pro. It does not promise a result: the Preview is a simulation.