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

# 🎙️ Voice Cloning & AI Voice Messages

> Clone your voice and send ultra-realistic voice messages from your conversations

The **voice cloning** feature lets you create a digital copy of your voice, then generate personalized audio messages directly from your conversations. Your leads will receive voice notes indistinguishable from real ones.

## Clone your voice

To clone your voice, go to **[Integrations](https://app.orsay.ai/integrations) → [Voice Cloning](https://app.orsay.ai/integrations/voice-cloning)**.

**1. Record audio samples**

Record yourself via the microphone or select one high-quality audio file.

* **One audio file at a time**
* Maximum **25 MB**
* Maximum **5 voices** per organization

**2. Name your voice**

Give your voice a name (e.g. "My pro voice", "Enthusiastic voice").

**3. Create the clone**

Click **"Clone"**. The voice is created in seconds and immediately available in your conversations.

## Generate a voice message

From any conversation, click the **audio** icon to open the voice message generator.

**1. Select a voice**

Choose from your cloned voices or the default voice.

**2. Write your message**

Type the text you want to convert to audio. The text will be automatically optimized for natural speech delivery.

**3. Choose the tone**

Three tones are available:

* **Professional** — stable and composed
* **Friendly** — casual and warm
* **Enthusiastic** — dynamic and energetic

**4. Advanced options (optional)**

* **Natural hesitations** — Add filler words (um, well, you know…) for an even more human feel. You can customize the word list.
* **Background noise** — Add ambient sound (Coffee shop or Cars) to make it sound like the message was recorded in a real setting.

**5. Preview and send**

Preview the audio, then click **"Use this audio"** to attach it to your WhatsApp message.

## Natural hesitations

To make your voice messages even more realistic, enable **natural hesitations**. The AI will inject filler words and self-corrections into your text before synthesis.

Default words: `um, well, so, like, you know, I mean`

> [!TIP]
>
> You can edit the hesitation word list directly in the text field. Adapt them to your speaking style!

## Background noise

Two ambient sounds are available:

| Ambiance | Description |
| - | - |
| ☕ **Coffee shop** | Coffee shop background — distant conversations, cups, warm atmosphere |
| 🚗 **Cars** | Traffic noise — cars, urban atmosphere |

The noise is mixed at a subtle volume so it doesn't overpower your voice.

## Manage your voices

From **[Integrations](https://app.orsay.ai/integrations) → [Voice Cloning](https://app.orsay.ai/integrations/voice-cloning)**, you can:

* View all your cloned voices
* Listen to a preview of each voice
* Delete a voice

> [!WARNING]
>
> Deleting a voice is permanent. Previously generated audio messages remain available, but you won't be able to create new messages with that voice.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.