How to clone your voice with AI?

Every video project runs on voices. A narrator for the cut, a character who needs one more take, a line you want read three different ways before the deadline.

Mitte’s Voice Cloner turns any of them into a voice you keep and reuse. Record 30 seconds, give it a name, and it’s ready to speak your script in two languages. Start to finish, about 2 minutes.

You build these up into a Voice Library that sits next to your projects, so a voice you cloned for one cut is right there for the next one.

Here’s how it works.

# Step 1: Open the Voice Cloner

Open the model menu, select Voice Library, then hit Voice Cloner.

That’s the whole way in. No keys, no separate app, no upload pipeline to wire up first.

# Step 2: Choose a voice to clone

Your own voice or one that already exists. You can do both. For this run I’ll use mine.

The flow has 3 parts. Snap a quick photo for the voice, then record a 30-second sample. Keep the sample under 90 seconds. A clean half-minute carries more signal than 3 rambling ones, and the model leans on that.

# Step 3: Give it a name

Once the photo and the sample are in, the only thing left is a name.

Type one into the prompt field. I’ll use Azer. You can describe the voice too if you want, though you don’t have to: the model writes its own name and description from what you hand it. Then press the blue button.

Give it a minute or 2. When it finishes, it drops you on the voice’s own page.

# Step 4: Make it speak

The voice is live. Paste the text you want spoken, set emotion, speed, and language if it matters for the line, then generate.

Type, generate, listen. That’s the loop.

# Step 5: Keep a library

You can build as many voices as a project needs and keep them together in your Voice Library.

Two ways to access them: the Presets page, and the model menu. Use whichever one matches how you work.

Clone your first voice at by clicking here.