Voicebox: an open-source local voice studio for cloning and dictation
Open-source local voice tool Voicebox combines instant voice cloning with system-wide dictation, racking up over 42,000 GitHub stars. It can clone voices from a few seconds of audio, type into any text field via global hotkeys, and voice coding agents like Claude Code. Basically, it lets you ditch paid voice subscriptions and run your entire workflow locally.
Voicebox is an open-source local voice studio designed to handle both voice cloning and speech-to-text right on your own machine. The project has already racked up over 42,000 stars on GitHub.
What it can do
Historically, running a full voice workflow meant paying for multiple tools: subscribing to ElevenLabs to make AI speak, and relying on apps like WisprFlow to turn your own speech into text. Voicebox aims to replace both ends in one free, local package.
On the input side, you can trigger speech dictation into any text box across your operating system using a global hotkey, giving you a smooth, hands-free typing workflow.
On the output side, feeding it just a few seconds of sample audio is enough to clone any voice. Beyond standard text-to-speech, it can integrate directly with coding agents like Claude Code and Codex, letting you assign custom cloned voices to your AI assistants.
Why it matters
If you lean heavily on voice tools, Voicebox can save you serious subscription cash. Running everything on your own machine cuts out per-character API costs and keeps sensitive audio entirely private.
Developers building agent workflows and power users who dictate all day should definitely give it a spin. Keep in mind that performance will depend directly on your local hardware specs, so you will want to benchmark latency against paid cloud alternatives.
Sources (2)
- github.com(opens in a new tab)Originalgithub.com
- voicebox.sh(opens in a new tab)Reportvoicebox.sh