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

# Languages

> Eleven, and both voices speak all of them.

| Code | Language       | Code | Language |
| :--- | :------------- | :--- | :------- |
| `hi` | Hindi          | `mr` | Marathi  |
| `bn` | Bengali        | `or` | Odia     |
| `gu` | Gujarati       | `pa` | Punjabi  |
| `kn` | Kannada        | `ta` | Tamil    |
| `ml` | Malayalam      | `te` | Telugu   |
| `en` | Indian English |      |          |

`en` is **Indian** English — English words in an Indian accent. It is not
British or American, and there is no code for those.

<Warning>
  **A wrong code is not an error.** `language` must match the script your text
  is actually written in. Hindi text with `"language":"ta"` is accepted and
  pronounced wrongly, with no error at all. Nothing warns you — you find out by
  listening.
</Warning>

## Code-mixed text

Leave `language` out when your text switches languages mid-sentence:

```json theme={null}
{ "text": "नमस्ते! Maya playground में आपका स्वागत है।" }
```

Auto-detection handles the mixture. Pinning a language forces every word through
one pronunciation model, including the ones that do not belong to it.