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

# Telnyx native voices on TTS

> List of Telnyx native text-to-speech models and voices, including supported languages, voice formats, latency tiers, and sample audio output.

## Models

| Model     | Voice Format                   | Latency | Languages                              |
| --------- | ------------------------------ | ------- | -------------------------------------- |
| Natural   | `Telnyx.Natural.<voice>`       | Low     | English                                |
| NaturalHD | `Telnyx.NaturalHD.<voice>`     | Medium  | en, fr, de, es, ar, hi, ja, he, pt     |
| KokoroTTS | `Telnyx.KokoroTTS.<voice>`     | Low     | —                                      |
| Qwen3TTS  | `Telnyx.Qwen3TTS.<clone_name>` | Medium  | en, zh, fr, de, it, ja, ko, pt, ru, es |
| Ultra     | `Telnyx.Ultra.<voice>`         | Medium  | Multilingual                           |
| Bayan     | `Telnyx.Bayan.<speaker>`       | Low     | Arabic (13 dialects) + English         |
| Sukhan    | `Telnyx.Sukhan.<voice_id>`     | Low     | Urdu                                   |

<Warning>
  **Ultra** is not available over WebSocket. Use the [REST API](/docs/voice/tts/rest-api) for Ultra.
</Warning>

## Natural & NaturalHD

Pre-built voices. Browse available voices via the [Voices API](/api-reference/text-to-speech-commands/list-available-voices) or the [Voice Design](https://portal.telnyx.com/#/app/ai/voice-design-lab).

### Audio Format

Default: MP3. NaturalHD supports `audio_format` query parameter to override:

```
?voice=Telnyx.NaturalHD.astra&audio_format=pcm
```

Accepted values: `pcm`, `wav`.

### Voice Settings

| Field             | Type    | Description                                            |
| ----------------- | ------- | ------------------------------------------------------ |
| `voice_speed`     | float   | Playback speed multiplier                              |
| `embedding_scale` | float   | Voice embedding intensity                              |
| `diffusion_steps` | integer | Quality/latency tradeoff — more steps = higher quality |
| `phonemizer`      | string  | Phonemizer backend selection                           |
| `response_format` | string  | Output format override                                 |
| `sampling_rate`   | integer | Sample rate in Hz                                      |
| `temperature`     | float   | Synthesis variability                                  |
| `volume`          | float   | Output volume                                          |
| `emotion`         | string  | Emotional tone                                         |

## Qwen3TTS

Voice cloning model. The `voice_id` is the name of a clone created in the [Voice Design](https://portal.telnyx.com/#/app/ai/voice-design-lab). Cloned voice usage may require identity verification.

Requires the clone to belong to your organization.

### Audio Format

Always raw PCM — 24kHz, signed 16-bit little-endian, mono. Forced by the backend regardless of any `output_format` value sent.

### Voice Settings

| Field            | Type    | Default  | Description                                                                                                                                                    |
| ---------------- | ------- | -------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `language_boost` | string  | `"Auto"` | Target language. Accepted: `Auto`, `English`, `Chinese`, `French`, `German`, `Italian`, `Japanese`, `Korean`, `Portuguese`, `Russian`, `Spanish`, or ISO codes |
| `force_xvector`  | boolean | `false`  | Force x-vector voice embedding                                                                                                                                 |

## KokoroTTS

Lightweight, low-latency model. Suitable for high-throughput applications where quality tradeoffs are acceptable.

## Bayan

Arabic voice model. 113 speakers across 13 dialects (Modern Standard Arabic, Egyptian, Emirati, Saudi, Jordanian, Iraqi, Lebanese, Syrian, Palestinian, Kuwaiti, Bahraini, Qatari, Omani) plus English speakers. Native audio is 16kHz.

### Voice Settings

| Field             | Type   | Default           | Description           |
| ----------------- | ------ | ----------------- | --------------------- |
| `language`        | string | Speaker's dialect | Language/dialect hint |
| `response_format` | string | native            | `pcm`, `wav`, `mp3`   |

Only the native `16000` sample rate is supported.

## Sukhan

Urdu voice model. 14 curated voices. No prosody controls or language selection — Urdu only. Native audio is 22050Hz.

### Voice Settings

None. `voice_speed` and other prosody settings are not supported and have no effect.

Only the native `22050` sample rate is supported. Supported `response_format`/`audio_format`: `pcm`, `mp3` (no `wav`).