🎙️ WAVESPEED AI VOICE GALLERY18 voices · 17 languages
All samples are original recordings made by WaveSpeed AI. Use them freely in your own videos, including commercial projects.
THREE SIMPLE STEPS
🎧

Listen and compare

Press play on any voice and switch the sample language to hear how it sounds in yours. Filter by style or voice type.

🎬

Use this voice

The AI Video Generator opens with the sample already attached as the reference audio and a starter prompt that points to it.

✍️

Write what they say

Describe the scene and put the words in quotes. The video is generated with a voice that follows your reference.

The sample goes in as reference audio. These models in the AI Video Generator accept it, each with its own limits.

Seedance 2.5 and 2.0

Up to 10 reference audios, 30 seconds in total. The default model when you open a sample from this page.

WAN 3.0

Up to 5 reference audios, 15 seconds in total.

MiniMax H3

Up to 3 reference audios. Audio cannot be the only reference: add an image or a video.

Kling 4.0

Up to 3 reference audios of 5 to 30 seconds each.

A voice is one part. Make the face, the moves and the video in the same place.

AI influencers and avatars

Give your character one consistent voice across every video instead of a new one each time.

Ads and explainers

Test an anchor, a friendly guide or a deep storyteller before you commit to a style.

Story and character videos

Match a narrator or a character voice to the scene, in the language of your audience.

How do I use a voice in a video?

Press Use this voice. The AI Video Generator opens with that sample attached as the reference audio and a starter prompt. Change the prompt to your scene and the words in quotes, then generate.

Will the video sound exactly like the sample?

The reference audio steers the voice of the speaker, so the pitch and character follow the sample. It is a guide for the model, not a copy, so results vary between models and prompts.

Which languages are the samples in?

Every voice speaks the same short line in 17 languages: English, Chinese (Simplified and Traditional), Japanese, Korean, Spanish, Portuguese, French, German, Italian, Russian, Turkish, Polish, Indonesian, Vietnamese, Thai and Hindi. Pick the sample language above the list.

Can I use the samples commercially?

Yes. They are original recordings made by WaveSpeed AI. You can use them in your own videos, including commercial projects, and download them for free.

How long are the samples?

About 9 seconds each (8 to 12 seconds). That fits the reference audio limits of Seedance, WAN 3.0, MiniMax H3 and Kling 4.0.

Do I need an account?

No account is needed to listen or download. Generating a video in the AI Video Generator needs a free account.

Your character, your voice, every video

Pick a voice, add your character, and generate a talking video in the AI Video Generator.

All Free Tools