yovoice

Audio & Video

Create speech and voiceovers locally from text, with model-dependent voice cloning and expression controls in a desktop workspace.

Use with conditions

Official website Repository

Last updatedRelease 2026-10-01

Open source licenseApache-2.0

PlatformsWindows 10/11 x64 · macOS 14+ on Apple Silicon

DeploymentDesktop

About this project

yovoice converts text into speech on a personal computer. Its desktop workspace supports reference recordings, voice cloning, and text-guided voice design, with expression controls that vary by model. The official site describes local generation after a model has been downloaded; its online audio examples play saved recordings rather than running a speech model in the browser.

The desktop app is distributed as a Windows x64 installer and an Apple Silicon DMG for macOS 14 or later. Users download a model in Settings, create a speech or story project, add text or subtitles, select a voice, and generate audio for preview and export. Windows supports CPU inference and optional NVIDIA CUDA, while Vulkan support is experimental.

Initial installation and model downloads require internet access. Memory requirements depend on the chosen model, and voices, recordings, and models are stored in ~/.yovoice by default; uninstalling does not remove that folder. The application uses Apache-2.0, but commercial audio use also depends on the model license and rights to the reference voice. OmniVoice weights are restricted to non-commercial use.

yovoice — site en yovoice — readme

Recommended by

GitHubDaily
@GitHub_Daily
Recommended on X
View original recommendation

Source labels distinguish recommendations, neutral mentions and the author’s own releases. Commercial relationships are unknown unless an author relationship is shown. Worthary reviews project eligibility separately.

Decision guide

Why it is worth a look

Local text-to-speech generation
Reference-voice cloning and voice design

Before you use it

Download a speech model before generation; the documented files range from hundreds of megabytes to several gigabytes
macOS requires Apple Silicon and macOS 14+; Windows supports CPU and optional NVIDIA CUDA, with Vulkan marked experimental
Commercial audio use depends on the selected model license and rights to the reference voice; OmniVoice weights are non-commercial
Voices and models remain in ~/.yovoice after uninstalling; backup and removal are separate tasks

Manual setupModerate

With an AI agentModerate

DeploymentDesktop

Additional requirementsNone required

A supported operating system and permission to install desktop software; Internet access and sufficient storage for the initial model download; Rights to any imported voice recording and a compatible model license

What it can do

Local text-to-speech generation
Reference-voice cloning and voice design
Model-dependent emotion controls
Audio preview and export

Who it fits

Good for

Creating narration or voiceovers on a supported personal computer; Reusing a reference voice for audio projects when the voice rights are clear

Not a fit

You need immediate speech generation without downloading models; Your project requires commercial use of a model whose weights are non-commercial; You require an Intel Mac desktop package

Getting started

Download the official Windows installer or Apple Silicon DMG and install the desktop app
Open Settings and download a speech model appropriate for your hardware and license needs
Create a speech or story project, enter text or import subtitles, and select a voice
Generate speech, preview the result, and export the audio

What it takes

Getting startedModerate

Ongoing maintenanceModerate

Failure impactModerate

Knowledge neededModerate

Not selected for comparison.

Get help from an AI agent

DeploymentDesktop

Additional requirementsNone required

With an AI agentModerate

{
  "project": "yovoice",
  "decision": "Use with conditions",
  "purpose": "Create speech and voiceovers locally from text, with model-dependent voice cloning and expression controls in a desktop workspace.",
  "repository": "https://github.com/leemysw/yovoice",
  "official_site": "https://yovoice.leemysw.com/en/",
  "deploy_shape": "Desktop",
  "additional_dependencies": "A supported operating system and permission to install desktop software; Internet access and sufficient storage for the initial model download; Rights to any imported voice recording and a compatible model license",
  "direct_setup": "Moderate: install the desktop app, choose and download a model, and check model and voice rights",
  "ai_assisted_setup": "Moderate: AI can explain model choices and platform setup; the user must check rights, storage, hardware, and generated audio",
  "operational_burden": "Moderate: keep the app and chosen models current, manage model storage, and back up local voice projects",
  "operational_risk": "Moderate: reference voices are sensitive recordings, local data survives uninstalling, and model licenses can constrain generated audio use",
  "knowledge_requirement": "Moderate: understand model selection, local storage, supported hardware, and permission to use a reference voice",
  "license": "Apache-2.0",
  "install_steps": [
    "Download the official Windows installer or Apple Silicon DMG and install the desktop app",
    "Open Settings and download a speech model appropriate for your hardware and license needs",
    "Create a speech or story project, enter text or import subtitles, and select a voice",
    "Generate speech, preview the result, and export the audio"
  ],
  "good_for": "Creating narration or voiceovers on a supported personal computer; Reusing a reference voice for audio projects when the voice rights are clear",
  "avoid_if": "You need immediate speech generation without downloading models; Your project requires commercial use of a model whose weights are non-commercial; You require an Intel Mac desktop package"
}

yovoice

About this project

Recommended by

Why it is worth a look

Before you use it

What it can do

Who it fits

Getting started

Alternatives to consider