Design AI voices from text descriptions or clone your own voice in 30 seconds — then use them across unlimited videos.

Every brand has a voice. Some sound corporate. Others sound friendly. A few sound unforgettable.

The Voices page is your voice laboratory — where you design AI voices from scratch or clone your own voice to use across every video GEN creates.

💡 Did you know?
71% of GEN users reduced manual editing time by 80% or more — and custom voices are a big reason why. No more recording the same script 20 different ways.

Click Voices in the left navigation to access your voice laboratory. Here you'll see all your created voices, their language settings, and creation dates. Each voice becomes available instantly in any Speech from Text creation card across all your projects.

Think of this as your voice casting department — except every "actor" is perfectly on-brand, available 24/7, and never needs a coffee break.

Maybe you need a voice that sounds "warm but authoritative" or "energetic Gen-Z." You don't need to find the perfect voice actor — you can design one.

From your Voices page, click Create Voice, then select Design a Voice.

Fill out the voice profile:

Your description might be: "Confident, slightly playful, like a smart friend explaining something cool. Think tech reviewer meets lifestyle influencer."

Add Sample Text — a sentence or two your voice will speak during generation. This lets you hear how it handles your actual content style.

Click Generate and GEN will create multiple AI voice samples matching your description. Listen to each option and select the one that nails your brand vibe.

⚡ Pro Move: Test your designed voice with the actual copy you'll use most often — product descriptions, calls-to-action, or trending hooks. The voice that sounds great reading generic text might not work for your specific content style.

Once you've selected your favorite sample, click Save. Your new voice appears in your library and becomes immediately available in all Speech from Text creation cards.

"I thought it would feel fake. It doesn't. My audience thinks I hired a team."
​— Founder, Supplement Brand

Want every video to sound exactly like you? Voice cloning creates a digital version of your voice that can speak any script while maintaining your natural cadence, accent, and personality.

Record yourself speaking for at least 30 seconds in a quiet environment. Read naturally — this isn't a robot training exercise. The better your source audio, the better your clone.

From your Voices page, click Create Voice, then select Clone Your Voice.

Fill out the same basic information as voice design (name, language, gender, description), then upload your MP3 file. The system needs at least 30 seconds but longer samples (up to 2-3 minutes) often produce better results.

Click Train Voice to start the AI training process. This typically takes 5-15 minutes depending on your audio length and current system load.

You'll get a notification when training completes. The system analyzes your vocal patterns, accent, speaking rhythm, and tonal qualities to create a digital version that can speak any text in your voice.

📈 How a DTC skincare brand uses this
The founder cloned her voice once, then used it across 52 videos in her first month. Her audience couldn't tell the difference, but she went from spending 3 hours a week recording to zero. "I sound like myself in every video, but I never have to speak."

Once you've created voices, they become available in every Speech from Text creation card across all your projects. Here's how the workflow connects:

In any vidsheet, when you select a cell and choose Speech from Text from the creation card menu, you'll see a Voice dropdown containing all your trained voices.

Choose your voice, add your script (or reference a variable from another column), and set timing preferences. Your custom voice will speak whatever text you provide — whether it's a static script or dynamically generated from AI text columns.

Click generate and your voice speaks the script. The audio file automatically appears in your timeline, perfectly timed and ready to layer with visuals.

💡 Did you know?
82% of GEN users said their audience couldn't tell the content was AI-generated — largely because they use consistent, branded voices that sound natural and on-brand.

Smart brands don't use one voice for everything. You might need different voices for different content types, products, or audience segments.

Name your voices descriptively so you can quickly select the right one:

⚡ Pro Move: Clone your voice multiple times with different emotional instructions in the description. Same voice, different energy levels. "Sarah - Calm" vs "Sarah - Excited" vs "Sarah - Urgent" gives you three voices from one recording session.

Don't settle for your first attempt. Create multiple versions, test them in actual videos, and see which ones your audience responds to best. You can always create new voices or refine existing ones based on performance.

💡 Did you know?
A supplement brand founder posted 31 videos in 21 days using voice cloning and saw an 18% increase in TikTok Shop revenue. The secret? Every video sounded like a personal recommendation from the founder, but required zero recording time.

With your voice library established, you're ready to create speech-driven content at scale. Your voices work seamlessly with GEN's other creation tools — combine them with AI-generated scripts, variables, and automated posting for a complete content engine.

Ready to start creating?

Browse our library of ready-made templates and launch your first video in minutes.

Browse Templates →