Yovoice

Yovoice is an open source desktop application for creating speech on your own computer. It covers text to speech, voice cloning, voice design from a written description, emotion control and multilingual voiceover. The software is released under the Apache 2.0 license and runs for macOS and Windows. All synthesis happens locally. Models are downloaded once, after which the app works offline and sends no requests to the cloud, so scripts and voice assets stay on the user's machine in a local folder. Users can clone a voice from a short reference recording, design a new voice from a text description, and steer delivery through emotion tags, reference audio or natural language prompts. The site demonstrates Mandarin, a Sichuan dialect, English, Japanese and French samples, along with laughter tags. Yovoice also provides a command line interface and an agent skill, so voiceover can be batch rendered from scripts or called by an AI agent without opening the graphical client. Supported models include IndexTTS 2.0 and 2.5, VoxCPM2, OmniVoice, Qwen3 TTS and Kokoro 82M, each with different strengths. Kokoro uses built in voices only and does not support cloning. The OmniVoice weights are limited to non commercial use, and commercial use of generated audio depends on each model's license. Hardware requirements are macOS 14 or later on Apple Silicon, or Windows 10 or 11 64 bit with CPU, NVIDIA CUDA or experimental Vulkan acceleration. The app is free to download and no paid plans are listed.