Voice Design

Describe a voice in your own words, and the AI builds it for you. Voice design lets you create completely unique voices that match your creative vision.

How It Works

1. Describe the voice you want in plain English

2. Generate — the AI creates a voice matching your description

3. Preview the result and adjust if needed

4. Save the voice to your library for reuse

Writing Good Descriptions

The more specific your description, the better the result. Include details about:

• Age — young adult, middle-aged, elderly

• Gender — male, female, androgynous

• Accent — British, Southern American, Australian, etc.

• Vocal texture — smooth, gravelly, breathy, crisp

• Personality — warm, authoritative, playful, mysterious

Examples

Per-Segment Emotion

Your designed voice has a consistent identity — the same fundamental sound across every line. But each individual segment can have its own emotion and delivery style.

The base voice stays the same while the performance changes:

• One line delivered as a whisper

• The next line shouted in anger

• A third line spoken with tender warmth

This gives you a consistent character voice with a full range of expression.

Improved Voice Consistency

Designed voices are built to keep the same recognizable speaker identity across lines and regenerations, while still letting each segment have its own emotion and delivery style.

Consistent and Reproducible

Designed voices are deterministic. The same description combined with the same seed value produces the same voice every time. This means you can reliably recreate a voice or share the exact settings with collaborators.

Saved to Your Library

Every voice you design is automatically saved to your personal library. You can:

• Reuse it across multiple projects

• Assign it to different characters

• Preview it at any time

• Add it to your voice picker for faster selection during audio generation

Availability

Voice design is available on paid plans (Starter, Pro, and Studio). Each plan includes a monthly voice design limit — check your plan details for specifics.

Input limits

• Voice name: up to 100 characters

• Voice description: up to 500 characters

• Sample text: up to 500 characters

50+ Languages

Designed voices support over 50 languages. Describe the voice in English, then generate speech in any supported language.