How long does it take to design a character voice?

About a minute. Type or pick a description, choose a preview line of up to 300 characters, and your character reads it as soon as the voice is ready.

How many credits does a character voice cost?

Designing a voice costs 500 credits, shown on the button before you start. Reading scripts with it is charged by length, like any other voice. See pricing for credit packs and plans.

Can the AI character voice generator imitate a celebrity?

It makes original characters only. A description that names a real person or a famous or copyrighted character is stopped by the safety check, and your credits stay with you. Describe the qualities you like and get a voice of your own.

Should I design a voice, use the library, or clone one?

Use the library when a ready-made voice fits; the studio suggests similar ones before you design. Design a voice when the character is new. To use your own voice, record a short sample with AI voice cloning.

Can I reuse my character voice on any script?

Yes. Every designed voice is saved in My voices, private to your account, with room for up to 20. Pick it in text to speech for any script, with tags and a pronunciation list.

Which languages and accents can a character voice have?

The accent menu covers nine English accents, such as American, British, Australian, Irish and Indian, plus Mandarin, Taiwan Mandarin and Cantonese. Your character's lines can be read in 36 languages.

Can I make male, female, child and elderly voices?

Yes. Choose female or male and an age from child, teen and young adult to middle-aged and senior, or write it in: "a curious eight-year-old girl" or "a wise wizard in his eighties".

How do I write a good character voice description?

Say who is speaking, how old they sound, the texture of the voice and how they talk. "A gruff pirate captain in his sixties with a deep, gravelly voice" works far better than "a pirate". You have up to 500 characters.

AI Character Voice Generator from Text

from description to a new voice: About 1 min
characters for your description: 500
preset voices included, in 24 languages: 300+
reuse your voice in text to speech: Any script
private voices per account: 20
credits per designed voice: 500

What is an AI character voice generator?

An AI character voice generator creates a new voice from a written description of a character: who they are, how old they sound, their accent and the way they talk. Instead of searching a list for something close, you describe the voice you hear in your head and use it on any script.

How to create a custom AI voice for your character

Three steps and about a minute of waiting.

Describe the character
Write who is speaking in up to 500 characters, or pick gender, age, accent, sound, pace and style from the menus. The full description builds as you go.
Listen to the preview
Choose a line of up to 300 characters. In about a minute your character reads it, and the voice is saved in My voices.
Voice any script
Open the voice in text to speech and give it a whole scene, with [laughs], [sighs] or [pause] for the acting.

What this character voice generator can do

Your character gets a voice made for it. Here the voice starts from your idea, and a ready-made cast waits when you need one now.

A custom AI voice from one description

Hunting a library for "close enough" can eat an afternoon. Write casting notes instead, such as "a gruff pirate captain in his sixties with a deep, gravelly, booming voice", and get a new voice with that age and texture in about a minute.

Build a voice from 6 quick picks

Stuck for words? Pick a gender, an age from child to senior, an accent, a sound such as warm, raspy or breathy, a pace and a style like villain or cartoon. The description writes itself, and you can edit every word.

Start now with 300+ preset voices

In a hurry? The studio first suggests preset voices that already sound close to your description. Each has a portrait and a name, in 24 languages, and the Characters filter gathers the ones made for roles.

Keep a cast of up to 20 private voices

A designed voice is saved in My voices, private to you, so episode twenty sounds like episode one. Direct each line with [laughs], [sighs] or [pause], and add invented names such as Zyrrak to the pronunciation list.

Ready-made character voices to use right now

Browse 300+ preset voices with portraits, filter by Characters and use any of them straight away, or design a new one for the role you need.

Character voice samples

Six voices designed from the description on their card, and six library voices playing a role. Send any description to the designer with one click.

Character voice generator guide: villains, creatures and anime

How to brief the AI character voice generator

The generator works like a casting director: the clearer the brief, the closer the first take. Cover five things and most descriptions land on the first try.

1Who: the role and its place in the story, such as a shopkeeper, a narrator or a final boss.
2Age and gender: "a teenage boy", "a woman in her forties", "a very old wizard".
3Texture: one or two sound words such as warm, raspy, breathy, bright or gravelly.
4Pace and energy: slow and deliberate, quick and excited, calm and steady.
5Attitude: kind, smug, nervous, playful or menacing, the feeling behind every line.

Pick from the menus first, then edit the sentence they build. It is the quickest way to a full brief, and every word stays yours to change.

Villain voices

Memorable villains rarely shout. Controlled menace comes from a slow pace, a low or smooth sound, and polite words with something cold underneath. Compare "a scary man" with "a cold, elegant villain in his fifties with a smooth, low voice, speaking slowly and precisely". The second gives the voice an age, a texture, a pace and an attitude to work from.

For a brute rather than a schemer, swap smooth for raspy and speed up the pace. Give the preview a line that sounds like an invitation and feels like a threat.

Fantasy creature and robot voices

Non-human characters still need a human anchor. Describe the creature, then the sound: "a tiny forest sprite with a light, airy, sparkling voice, quick and giggly", or "a friendly service robot with a bright, slightly metallic, evenly paced voice". Words like metallic, airy or rumbling set the texture.

Think about size and age, too: an ancient dragon might be "very old, deep and slow, with a rumbling voice and long pauses". Keep its lines in plain language and let the voice carry the strangeness.

Anime and cartoon voices

Anime voices are expressive and a little larger than life, so name the archetype and the energy: "a cheerful teenage girl with a bright, high, quick voice". Cartoon voices push one quality further: "a tiny, excitable mouse with a squeaky voice who talks fast".

Write the big moments into the script as well. A [pause] before a reveal or a punchline, and a [laughs] after a victory line, let the voice react the way the genre expects.

Cast a whole story

Games, audio dramas and animated series need voices that are easy to tell apart. Vary at least two of age, pitch, pace and accent between characters who share a scene: a quick, bright young hero next to a slow, deep mentor is always clear.

Keep a cast sheet with each character’s description and preview line. When the story adds a new face, the sheet shows which sounds are taken, and a new custom AI voice fills the gap in about a minute.

Choose a preview line that shows the character

The preview line is the audition. Choose a line of up to 300 characters that shows attitude as well as information: "You came all this way for a map? How adorable." tells you far more about a villain than "Welcome to the castle".

Add one tag where the character would react, a [laughs] or a [pause], and listen to how the voice carries it. When the feeling is right in one line, it carries a whole scene.

Create AI voice variations and keep the best

When a voice is close, change one word in the description and create it again: raspy to smooth, slow to measured, fifties to seventies. Two versions side by side make the right one obvious.

Name each voice after its character, such as Captain Rook, and it waits in My voices, ready for text to speech. Episode twenty then sounds exactly like episode one.