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

# Voice Lab

> Create custom hybrid voices by blending multiple Aero TTS voices together.

# Voice Lab

Voice Lab lets you create custom hybrid voices by blending 2-4 existing Aero TTS voices together. Adjust the weight of each voice to craft a unique sound that matches your brand.

## How Voice Mixing Works

When you create a voice mix, Aero TTS blends the acoustic characteristics of multiple voices based on the weights you assign. Higher weight means more influence from that voice.

```
Voice Mix = Voice A (weight 1.5) + Voice B (weight 0.8)
```

The result is a new voice that combines qualities from both — for example, the warmth of one voice with the clarity of another.

## Creating a Voice Mix

### In the Dashboard

1. Go to **Aero TTS** > **Voice Lab**.
2. Click **Create Mix**.
3. Select 2-4 base voices.
4. Adjust the weight slider for each voice (0.1 to 10.0).
5. Enter a name and optional description.
6. Click **Preview** to hear the result.
7. Click **Save** to create the mix.

Your mix appears as a custom voice in the playground and API.

### Via API

```bash theme={null}
curl -X POST https://api.thinnest.ai/api/tts/voice-mixes \
  -H "Authorization: Bearer thns_sk_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Warm Professional",
    "voices": [
      { "voice": "aero-vayu", "weight": 1.5 },
      { "voice": "aero-aria", "weight": 0.8 }
    ],
    "description": "Blend of Vayu warmth with Aria friendliness"
  }'
```

**Response:**

```json theme={null}
{
  "voice_id": "mix_abc123",
  "name": "Warm Professional",
  "mix_config": [
    { "voice": "aero-vayu", "weight": 1.5 },
    { "voice": "aero-aria", "weight": 0.8 }
  ]
}
```

## Using a Voice Mix

Once created, use the mix `voice_id` in any synthesis request:

```bash theme={null}
curl -X POST https://api.thinnest.ai/api/tts/synthesize \
  -H "Authorization: Bearer thns_sk_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "This is my custom blended voice.",
    "voice": "mix_abc123",
    "model": "aero",
    "format": "mp3"
  }' \
  --output custom-voice.mp3
```

You can also pass `voice_mix` inline without saving a mix:

```bash theme={null}
curl -X POST https://api.thinnest.ai/api/tts/synthesize \
  -H "Authorization: Bearer thns_sk_YOUR_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "text": "Inline voice mixing without saving.",
    "voice": "aero-vayu",
    "model": "aero",
    "voice_mix": [
      { "voice": "aero-vayu", "weight": 1.5 },
      { "voice": "aero-kian", "weight": 0.5 }
    ],
    "format": "mp3"
  }' \
  --output inline-mix.mp3
```

## Mix Parameters

| Parameter | Type   | Range      | Description                        |
| --------- | ------ | ---------- | ---------------------------------- |
| `voice`   | string | —          | Voice ID of a base voice           |
| `weight`  | number | 0.1 - 10.0 | Influence of this voice in the mix |

### Weight Guidelines

| Weight    | Effect                                                  |
| --------- | ------------------------------------------------------- |
| 0.1 - 0.5 | Subtle influence — adds undertones                      |
| 0.5 - 1.0 | Moderate — blended equally                              |
| 1.0 - 2.0 | Dominant — primary voice characteristics                |
| 2.0+      | Heavy dominance — the mix sounds mostly like this voice |

## Tips for Great Mixes

* **Start with 2 voices** — More voices make results less predictable.
* **Use similar languages** — Mixing `en-us` and `en-gb` voices works well. Cross-language mixing may produce unexpected results.
* **Balance weights** — A 1.5:0.8 ratio sounds natural. Extreme ratios (10:0.1) essentially use one voice.
* **Test with varied text** — A mix that sounds good for short phrases may behave differently with long paragraphs.
* **Name descriptively** — "Warm Professional" is more useful than "Mix 1".

## Managing Mixes

### List Your Mixes

```bash theme={null}
curl https://api.thinnest.ai/api/tts/voice-mixes \
  -H "Authorization: Bearer thns_sk_YOUR_KEY"
```

### Delete a Mix

```bash theme={null}
curl -X DELETE https://api.thinnest.ai/api/tts/voice-mixes/mix_abc123 \
  -H "Authorization: Bearer thns_sk_YOUR_KEY"
```

Deleting a mix does not affect past audio generated with it. Future synthesis requests using the deleted mix ID will fail.

## Limits

* Maximum **4 voices** per mix
* Minimum **2 voices** per mix
* Mix name must be 1-100 characters
* Mixes are per-user — not shared across team members