We're upgradingWe ship improvements every week. Hit a problem, or have an idea? We reward good suggestions.Talk to us
Multi-speaker projects

Write a conversation. Cast the voices. Export one file.

Voice Studio is where a script with more than one speaker becomes finished audio. Give each character its own voice, direct individual lines, and export the whole project as a single track or as numbered clips.

How Voice Studio works

Four steps from a blank project to an exported dialogue.

  1. 1

    Create a project

    Every dialogue lives in its own project, saved as you work. Come back to it any time.

  2. 2

    Cast your speakers

    Add a speaker for each character and assign them a voice from the library — a stock voice, a voice you cloned, or one you designed.

  3. 3

    Write the script line by line

    Each line belongs to a speaker. Reorder freely, and group long scripts into chapters so a 200-line script stays navigable.

  4. 4

    Generate and export

    Generate a single line while you iterate, or generate everything at once. Export as one combined file or as separate numbered clips.

Built for scripts, not one-off clips

The difference between a text-to-speech box and a project editor.

A cast, not a voice

Assign a distinct voice to each speaker once, then every line they say uses it. Change a speaker's voice and their lines re-cast together.

Regenerate one line

Not happy with line 47? Regenerate just that line. The other 199 stay exactly as they were.

Chapters

Split a long script into named chapters — episodes, scenes, or modules — and move lines between them.

Every take is kept

Each line keeps its generation history, so you can compare takes and restore an earlier one instead of chasing a result you already had.

Direct the performance

On the Expressive model, steering lets you say how a line should be delivered — pace, pitch, intensity — instead of only what it says.

Three models, three budgets

Swift at half credits for rough drafts, Ultra HD as the studio default, Expressive when a line has to act.

Export the way you edit

One combined MP3, a lossless combined WAV, or a .zip of one numbered MP3 per line to drop straight onto a timeline.

16 languages, Turkish included

Write and generate in Turkish, English, Spanish, German, French and 11 more — with the same cast across the whole project.

What people build with it

Anything where more than one person speaks.

Podcasts and interviews

A two-host format, scripted and generated, without booking two people and a room.

E-learning and courses

A narrator plus a student voice makes a module far easier to follow than a single track of narration.

Ads and promos

Script a 30-second spot with a voice-over and a character, then export a single mixed file.

Audiobooks and fiction

Give the narrator one voice and each character another, chapter by chapter.

Animation and games

Generate placeholder or final dialogue for every character from one script.

Explainers and demos

A question-and-answer format keeps a product walkthrough from sounding like a manual.

Frequently asked questions

?How many speakers can one project have?

As many as your script needs. Each speaker is a name plus a voice, and every line is assigned to one of them.

?Can I use a voice I cloned or designed?

Yes. Any voice in your library — stock, cloned, or designed — can be cast as a speaker.

?Do I have to regenerate everything after an edit?

No. Editing a line's text or speaker resets only that line. Structural changes such as reordering keep the audio you already generated.

?Does Voice Studio work in Turkish?

Yes. Turkish is one of the 16 output languages, and a project can keep the same cast across every line.

?What can I export?

A combined MP3, a lossless combined WAV, or a .zip containing one numbered MP3 per line. Combined exports need every line generated first.

?What does it cost?

Generation is charged in credits by character count, and the model sets the multiplier — Swift 0.5×, Ultra HD 1×, Expressive 3×. Editing and organising a project is free.

Your script already has more than one voice in it

Create a project, cast the speakers, and hear the first line in a couple of minutes.

Start in seconds

No credit card needed

Start free