yovoice
Download yovoice

Give your words a voice.

Choose a voice, set the mood, and hear your words come to life.

Long-form performance / Multiple characters3 models01 / 14 80

他脸上黑而且瘦，已经不成样子；穿一件破夹袄，盘着两腿，下面垫一个蒲包，用草绳在肩上挂住；见了我，又说道，“温一碗酒。”掌柜也伸出头去，一面说，“孔乙己么？你还欠十九个钱呢！”孔乙己很颓唐的仰面答道，“这……下回还清罢。这一回是现钱，酒要好。”掌柜仍然同平常一样，笑着对他说，“孔乙己，你又偷了东西了！”但他这回却不十分分辩，单说了一句“不要取笑！”“取笑？要是不偷，怎么会打断腿？”孔乙己低声说道，“跌断，跌，跌……”他的眼色，很像恳求掌柜，不要再提。此时已经聚集了几个人，便和掌柜都笑了。我温了酒，端出去，放在门槛上。他从破衣袋里摸出四文大钱，放在我手里，见他满手是泥，原来他便用这手走来的。不一会，他喝完酒，便又在旁人的说笑声中，坐着用这手慢慢走去了。

00:00 / 01:11Ready
yovoice

Swipe to explore the liner notes →

VOICE COLLECTION / 001
Listening collection · Vol. 0112 samples / 04 ways to create
CATALOG
YV-001
TRACKS
01—12

DEMO Listen freely

SIDE A
SIDE A
SIDE B
SIDE B

Create on your own computer.

Readings · Character performances · Voiceovers

  1. Write your script

    From a greeting to a whole story

  2. Choose a voice

    Pick a voice or reuse a familiar one

  3. Shape the emotion

    Match the delivery to your words and listen

The yovoice workspace with a text editor, voice and emotion settings, and an audio timeline
WORKSPACE / 01Your local voice workspace

Find the voice. Shape the delivery.

A retro yovoice cassette A few seconds of audio. A familiar voice.
01 / Voice

Clone a voice. Or create one.

Upload or record a reference clip to clone a voice, or describe the voice you have in mind to create a new one.

YOVOICE / EXPRESSION
Same words. A different feeling.

你终于来了，我等你好久了。

Switch the mood and listen to what changes.
02 / Expression

Let the delivery do the talking.

Bright and excited, or quiet and restrained? Guide the delivery with emotion tags, a reference recording, or a written prompt.

Available controls vary by model.
A retro desktop computer showing a waveform and a saved voice.wav file
03 / On your computer

Your audio stays on your computer.

Download a model once, then generate audio locally. Your scripts and voices stay on your computer, without a cloud service.

Available for macOS and Windows
04 / Your workflow

Make voice generation part of your workflow.

Generate audio from terminal scripts or let an agent call yovoice. No need to open the desktop app for each task.

Agent Skill · CLI
Supported models Model capabilities and licenses
Models supported by yovoice and their uses
Model What it does
IndexTTS 2.0 / 2.5 Voice cloning, emotion control, and reference-guided delivery. Version 2.5 also supports multiple languages and pronunciation adjustments.
VoxCPM2 Design a voice with text, or clone one using a reference recording and its transcript.
OmniVoice Voice design, voice cloning, and non-speech tags. Model weights are for non-commercial use only.
Qwen3-TTS Base clones a reference voice, CustomVoice offers preset voices, and VoiceDesign creates voices from text descriptions.

FAQ

How were these audio samples made?

The samples were generated locally with the yovoice CLI. Each sample names the model used. The calm and happy examples share the same text and reference voice, with different emotion settings. This page plays saved audio; it does not run a model in your browser.

Where are my voices, recordings, and models stored?

By default, they are stored in ~/.yovoice on your computer. Uninstalling the app does not remove this folder. Back it up when moving to another computer.

Do I need an internet connection?

You can generate audio offline.An internet connection is needed to install the app and download models. Once a model is ready, speech synthesis runs on your computer.

What are the system requirements?

Two platforms are supported:

  • macOS: macOS 14 or later on Apple Silicon, with native Metal acceleration.
  • Windows: Windows 10/11 (64-bit), with CPU, NVIDIA CUDA, and experimental Vulkan support.

RAM and VRAM requirements depend on the model you choose.

Can I use the audio commercially?

yovoice is licensed under Apache-2.0. Commercial use also depends on the model license and your rights to the voice:

  1. Each model has its own license. For example, OmniVoice weights are restricted to non-commercial use.
  2. Before commercial use, check that the model allows it and that you have the necessary rights to the reference audio and voice.
Can an AI agent generate audio with it?

Yes. yovoice includes a standalone CLI and an Agent Skill. Scripts and agents can generate audio from text and a reference voice without opening the desktop app.

View the Agent Skill

yovoice DESKTOP APPLICATION

Your next voice starts here.

Local generation · Emotion control · Voice cloning

Get yovoiceOpen GitHub Releases
macOS 14+Apple SiliconWindows 10/11x64