voicestudio

VoiceStudio is an open-source, fully local alternative to cloud voice services, covering voice cloning, voice design, video dubbing, dictation, transcription and audiobook creation. The project advertises support for 646 languages and positions local workflows as running on the user's own hardware, with remote services optional and usage analytics requiring consent. Core capabilities are grouped into three areas. Create: clone a voice from a reference recording or design a new voice from a description. Produce: dub videos with timed speech, generate stories and audiobooks, and run batch jobs. Connect: expose a local API and MCP server for agents, with optional remote workers. The default engine is VoiceStudio, described as powered by k2-fsa/OmniVoice, and other engines can be selected; a feature and engine catalog is provided in the docs. Hardware requirements vary by engine, and a performance document covers expectations. Models are downloaded on demand, and the app prompts before installing required models. Installation options include a one-command installer for macOS and Linux via a curl script, which supports installing the latest Electron release, a specific version, building current main, or uninstalling while keeping data. Release downloads require curl and a SHA-256 tool; building main requires Git, Node.js 22+, Bun, Rust/Cargo and platform build tools. Platform guides exist for macOS, Windows, Linux and Docker, and releases can be downloaded directly. There is also an agent-oriented install prompt for coding agents such as Claude Code, Codex and Cursor, plus an agent guide covering hardware detection, reusing existing data, asking before model downloads and a test generation. Skills can be added with npx skills add debpalash/VoiceStudio, choosing either the audio workflow skill or the maintainer skill. Running from source uses Bun: clone the repository, run bun install, bun run setup:api to prepare Python dependencies, then bun run dev. A smoke-test command builds and launches an isolated packaged Electron app, with an optional networked managed-runtime installation check. The README states that Electron is the only maintained desktop app and that version 0.5.3 was the final Tauri release; existing Tauri users are directed to a migration document, and Tauri source is archived without further updates. Root development, build, test and release commands target Electron. Documentation covers troubleshooting, model downloads, engine guides, benchmarks, the local speech platform API, MCP integration and examples, plus contributing, Electron setup and a changelog. The project is licensed AGPL-3.0, with a separate notice covering model licenses and responsible use; the README advises reviewing model licenses before commercial use and cloning voices only with permission. Sponsorship and donation channels are listed.