VoiceStudio

Overview

GitHubGitHub

VoiceStudio is an open-source desktop app built on Electron, positioned as a local alternative to ElevenLabs. It combines voice cloning, voice design, video dubbing, dictation, transcription, and audiobook creation in a single interface, supporting 646 languages. The app uses the k2-fsa/OmniVoice engine by default but allows users to choose other engines. All workflows run on local hardware; remote services are optional, and analytics require explicit user consent. **Key capabilities include:** - **Voice cloning & design**: Clone voices by uploading reference recordings, or generate custom voices from text descriptions. - **Video dubbing**: Add timeline-aligned voiceovers to videos. - **Dictation & transcription**: Provides a floating widget for voice dictation and supports speech-to-text. - **Batch processing**: Supports batch generation tasks for stories and audiobooks. - **Integrations**: Offers local API and MCP interfaces for use with AI agents (such as Claude Code, Codex, Cursor). - **Model management**: Built-in model manager for installing and managing local voice models, displaying license and review information. **Hardware compatibility:** - NVIDIA GPUs (Windows/Linux) support CUDA acceleration. - Apple Silicon supports Metal (MPS) acceleration. - Devices without a discrete GPU can run on CPU, though slower, requiring about 5 GB of disk space to install the CPU version of PyTorch. - Windows on ARM is experimentally supported. - Intel Macs run only the UI and must connect to a remote backend. VoiceStudio is open-sourced under AGPL-3.0, with installers available as macOS `.dmg`, Windows `.exe`, and Linux `.AppImage` or `.deb`. Building from source requires Git, Node.js 22+, Bun, Rust/Cargo, and platform build tools. Note that Electron is currently the only desktop app and Web UI format; the legacy Tauri architecture has been removed.

Q&A

Which operating systems does VoiceStudio support?AI-generated

Supports macOS (.dmg), Windows (.exe), and Linux (.AppImage or .deb). Electron is currently the only desktop app format; the legacy Tauri architecture is no longer maintained and has been removed.

Can it run without a discrete GPU?AI-generated

Yes. Without a discrete GPU (e.g., Intel/AMD integrated graphics or older PCs), the app runs entirely on CPU, though more slowly. The installer automatically downloads the CPU version of PyTorch, which is about 5 GB.

Is VoiceStudio free?AI-generated

VoiceStudio is open-source software under the AGPL-3.0 license and is free to use, including commercially when license terms are met. However, licenses for model weights, tokenizers, voices, and generated content are separate.

How do I build VoiceStudio from source?AI-generated

You need Git, Node.js 22+, Bun, Rust/Cargo, and platform build tools installed. After cloning the repository, run bun install, bun run setup:api, and bun run dev in sequence to start the development environment.

Can I use it on an Intel Mac?AI-generated

Intel Macs can only run the UI portion of the app and cannot run the backend locally. You must connect to a remote backend to access full functionality.

What voice engine does it use?AI-generated

It uses the k2-fsa/OmniVoice engine by default, but users can also choose other engines. Refer to the official documentation for the engine catalog and feature mapping.