yovoice Lets You Clone Voices Locally Without Paying ElevenLabs
A new open-source desktop app bundles IndexTTS, VoxCPM2, OmniVoice, and Qwen3-TTS for local voice cloning on Mac and Windows, with no cloud fees.
- yovoice is an open-source desktop TTS app for macOS and Windows with no cloud fees.
- Bundles IndexTTS 2.0/2.5, VoxCPM2, OmniVoice, and three Qwen3-TTS variants in one UI.
- Supports voice cloning from 1 to 60 second reference clips, plus text-prompt voice design.
- Runs on Metal, CUDA, CPU, and experimental Vulkan via the audio.cpp inference engine.
- Ships an Agent Skill so AI agents can install the CLI and generate speech autonomously.
- Apache-2.0 licensed app; OmniVoice weights restricted to non-commercial use.
Hosted services such as ElevenLabs charge by character, which can make long-form narration and repeated takes expensive. yovoice repository packages several recent text-to-speech models in a desktop app for macOS and Windows, with synthesis running on the user’s hardware. The application code is available under the Apache License 2.0.
Once a model is downloaded, generation avoids cloud round trips and metered API fees. Reference recordings and generated speech remain on the machine during inference. The project site lists voice cloning, emotion control, text-guided voice design, and project management among the app’s core features.
Several models, one workbench
yovoice wraps audio.cpp, an inference runtime that executes neural speech models on desktop hardware. Its model picker supports resumable downloads and imported GGUF files, a model format commonly used by local inference tools.
| Model | Capabilities in yovoice |
|---|---|
| IndexTTS 2.0 | Chinese and English voice cloning, emotion control, and delivery based on a reference performance. |
This story is for Pro members
You've reached the end of the free preview. Upgrade to AlphaSignal Pro to read the full article - and everything else behind the paywall.
Comments
