# Voicebox

A local-first voice studio that clones voices, generates speech in 23 languages, and dictates into any app.

HTML: https://softbests.com/software/voicebox
Markdown: https://softbests.com/software/voicebox.md
Website: https://voicebox.sh

- Category: [Voice input](https://softbests.com/category/voice-input)
- Pricing: Open source
- Platforms: macOS, Windows, Linux
- Underrated

## Best for

Creators and agent users who want ElevenLabs-style output and Wispr Flow-style dictation without sending voice data to the cloud.

## Why it deserves attention

Voicebox combines voice cloning, 8 TTS engines, 50+ preset voices, 23 languages, audio effects, a stories editor, and a global dictation hotkey in one native Tauri app. Everything runs locally, including Whisper-based speech recognition and a bundled LLM for rewriting or persona responses, and an MCP server lets Claude Code, Cursor, and Cline speak through a cloned voice. It is MIT-licensed with 59,000 GitHub stars and free, with macOS, Windows, and Docker builds. Local inference needs a capable machine, Linux binaries are not prebuilt yet, and cloning a voice carries consent and misuse risks that the app cannot enforce.

## Can replace

- ElevenLabs

## Related and alternatives

- [Wispr Flow](https://softbests.com/software/wispr-flow)
- [Superwhisper](https://softbests.com/software/superwhisper)