> Fetch clean Markdown by appending `.md` to any page URL under https://signalwire.com/docs or requesting it with the HTTP header `Accept: text/markdown`. The root index at https://signalwire.com/docs/llms.txt lists the available documentation indexes.

# MiniMax

> Learn how to use MiniMax's expressive, multilingual TTS voices with SignalWire AI Voice applications.

MiniMax is a text-to-speech engine offering expressive voices across many languages, with controls
for emotion, speed, pitch, and volume.

## Models

SignalWire supports the following MiniMax models. Pick a `turbo` model for speed and cost, or an
`hd` model for the highest quality. If you don't specify a model, `speech-2.6-turbo` is used.

| Model              | Description                                            |
| ------------------ | ------------------------------------------------------ |
| `speech-2.6-turbo` | **Default.** Latest turbo model: faster and lower-cost |
| `speech-2.6-hd`    | Latest HD model: higher quality                        |
| `speech-02-turbo`  | Previous-generation turbo model                        |
| `speech-02-hd`     | Previous-generation HD model                           |
| `speech-01-turbo`  | First-generation turbo model                           |
| `speech-01-hd`     | First-generation HD model                              |

## Voices

MiniMax provides a large library of system voices across many languages.
A voice's **Voice ID** is what you put in the [voice string](#usage): for example,
`English_CalmWoman` becomes `minimax.English_CalmWoman`.

> **Warning**
>
> Only MiniMax's built-in **system voices** are supported. Cloned and AI-generated voices do not
> work — you must use one of the default voice IDs that MiniMax provides.

> **Info**
>
> Some voice IDs contain spaces or parentheses (for example, `Cantonese_ProfessionalHost (F)`).
> Wrap the whole voice string in quotes when a voice ID isn't a plain word, for example
> `voice: "minimax.Cantonese_ProfessionalHost (F):speech-2.6-turbo"`.

Browse the full list of voice IDs in MiniMax's documentation:

#### [MiniMax system voice IDs](https://platform.minimax.io/docs/faq/system-voice-id)

Browse the full list of MiniMax system voice IDs.

#### [MiniMax TTS docs](https://platform.minimax.io/docs/api-reference/speech-t2a-http)

Reference for models, voices, and parameters.

## Languages

MiniMax supports the following languages and automatically detects the language of your text:

Arabic, Cantonese, Chinese — Mandarin, Czech, Dutch, English, Finnish, French, German, Greek, Hindi, Indonesian, Italian, Japanese, Korean, Polish, Portuguese, Romanian, Russian, Spanish, Thai, Turkish, Ukrainian, Vietnamese.

Refer to the [MiniMax developer platform](https://platform.minimax.io/)
for the most up-to-date reference to supported languages and voices.

## Usage

A voice identifier string has three parts: the `minimax` engine code, a voice ID, and an optional model.

**Format**: `minimax.<voiceId>:<model>`

* `voiceId` (required): a MiniMax system voice ID (for example, `English_CalmWoman`); see [Voices](#voices).
* `model` (optional): one of the [models](#models) above. Defaults to `speech-2.6-turbo`.

**Examples**:

```text
minimax.English_CalmWoman:speech-2.6-turbo
minimax.English_Trustworth_Man:speech-2.6-hd
minimax.Japanese_KindLady:speech-2.6-turbo
```

You can also set the model with the separate
[`model`](/docs/swml/reference/ai/languages#use-voice-strings)
parameter of the [`languages`](/docs/swml/reference/ai/languages) SWML method
instead of appending `:model` to the voice string:

```yaml
languages:
- name: English
  code: en-US
  voice: minimax.English_CalmWoman
  model: speech-2.6-turbo
```

---

## Build with MiniMax on SignalWire

### Create a Space and add credit

If you don't have one yet, you'll need to
[create a SignalWire Space](/docs/platform/signing-up-for-a-space).
Be sure to add some credit to test with.

### Add a new Resource

Navigate to the **Resources** tab in your [SignalWire Dashboard](https://my.signalwire.com) and click **+ Add New** to create a new Resource.

### Create a SWML Script

From the Resources menu, select **SWML Script**.
Name it something fun and recognizable.
Ours is titled MiniMax Wizard.

Next, paste the following starter script into the text box, and hit <kbd>Save</kbd>:

```yaml
version: 1.0.0
sections:
  main:
  - ai:
      prompt:
        text: |
          You're the Calm Woman, a voice from MiniMax's TTS engine!
          Introduce yourself, and have a conversation about programmable unified communications on the SignalWire platform.
      languages:
      - name: English
        code: en-US
        voice: minimax.English_CalmWoman:speech-2.6-turbo
```

### Buy and assign a phone number

Navigate to the **Phone Numbers** section of the Dashboard's left sidebar menu.

Purchase a phone number and assign it to the desired SWML script.

In the Dashboard, open **Phone Numbers** > **Purchased**, select the number, and open its edit page. Under **Inbound Call Settings**, select **+ Assign Resource**, choose the SWML script, and save the settings.

### Give it a call!

Call the number you just assigned to chat with your new AI voice application on the phone.

## Next steps with SWML

Now you've deployed your very first SignalWire voice AI application using MiniMax voices.
Next, dive deeper into SWML to explore its capabilities!

#### [Methods reference](/docs/swml)

Documentation for all SWML methods

#### [AI in SWML](/docs/swml/guides)

Build advanced AI applications using SignalWire Markup Language

#### [Guides](/docs/swml/guides)

SWML guides and demo applications

> Text-to-speech