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

# Voices

> Studio-grade voices from five TTS providers - pick a voice from the library, tune pronunciation, and set fallbacks.

Clara speaks with real-time text-to-speech. Choose a voice from the voice library - 94 stock voices across five providers, with copyable voice IDs and audio previews in your dashboard.

## Voice providers

| Provider       | Models               | Best for                                           |
| -------------- | -------------------- | -------------------------------------------------- |
| **ElevenLabs** | Turbo v2.5, Turbo 2  | Low latency, fast multilingual, the default choice |
| **Cartesia**   | Sonic 3.5, Sonic 3   | Next-gen real-time streaming voices                |
| **Rime**       | Mist v3, Coda        | Enterprise clarity at low latency                  |
| **Fish Audio** | S2 Pro, S2.1 Pro     | State-of-the-art expressiveness                    |
| **Inworld**    | Inworld TTS 1.5 Mini | Character-driven, interactive voices               |

## Choosing a voice

| You want…                     | Look at                                     |
| ----------------------------- | ------------------------------------------- |
| Professional, corporate tone  | Chris, Will, Evan (ElevenLabs)              |
| Warm, conversational delivery | Rachel, Laura, Hope (ElevenLabs)            |
| Real-time streaming           | Brooke, Blake, Katie (Cartesia)             |
| Enterprise clarity            | Luna, Celeste, Orion (Rime)                 |
| Expressive, natural delivery  | Hannah, Adrian, Ethan (Fish Audio)          |
| Broad multilingual coverage   | ElevenLabs Turbo v2.5 or Cartesia Sonic 3.5 |

## Setting a voice

1. Open the agent's settings and go to the **Voice** section.
2. Browse the **Voice Library** - play previews and copy the voice ID.
3. Select the voice and save. It applies on the next conversation.

## Voice settings

Fine-tune how Clara speaks (defaults apply if left unset):

| Setting          | What it controls                                    |
| ---------------- | --------------------------------------------------- |
| Voice ID         | The voice Clara uses                                |
| Language         | Spoken language output                              |
| Speaking rate    | Pace of delivery                                    |
| Stability        | Consistency vs expressiveness                       |
| Similarity boost | Closeness to the original voice character           |
| Fallback voice   | Backup voice if the primary provider is unavailable |

## Custom pronunciations

Teach Clara how to say your words - product names, acronyms, jargon:

| Written as | Spoken as   |
| ---------- | ----------- |
| `API`      | A P I       |
| `SQL`      | Sequel      |
| `ClaraSDR` | Clara S D R |

Add pronunciations per agent in the voice settings.

## Voice cloning

Clone a brand-consistent voice from professional voice samples through ElevenLabs Professional Voice Cloning - [contact us](mailto:support@trugen.ai) to enable it, or pass an existing cloned voice ID when configuring the agent. See [pricing on clarasdr.ai](https://clarasdr.ai/pricing) for plan availability.