Faster, clearer, more reliable local voice production

This release focuses on the parts that matter in everyday work: complete audio, honest engine status, smoother editing, and safer local workflows.

What changed

Designed voices work across engines

Saved Voice Design voices can now run with Fast, Expressive, Multilingual, or Omni when the required model is installed.

Long Omni scripts fail clearly instead of returning cut-off audio

Vois also tells you when delivery speed was adjusted.

Numbers sound right

Prices, decimals, grouped amounts, and version numbers are normalized correctly before synthesis.

Faster app, lower memory use

Startup is quicker, long lists and timelines render more smoothly, caches have sensible limits, and idle models unload automatically.

Safer generation

Vois checks the selected voices, engines, models, and licence before starting a batch. If a run cannot start, existing timeline audio stays intact.

Clearer feedback

Model status now distinguishes installed, missing, and unreadable packages. If Vois uses another engine for the selected language, it tells you which one ran.

More reliable exports and automation

Preview and export audio now use the same processing path, concurrent exports no longer collide, and repeated Windows CLI commands work for the life of the app.

Licensing and privacy hardening

Activation and offline-state handling are clearer, AppSumo lifetime licence flows are more reliable, local app control requires authentication, and sensitive data is kept out of logs.

Download Vois 2.1.6

Available now for Apple Silicon Macs and Windows.

Download Vois