FAQ

What is VoiceWiz?

A desktop voice-to-work application. VoiceWiz combines transcription, context, and optional refinement to turn spoken language into text that fits the task — code instructions, prose, research notes, or emails.

Is VoiceWiz just a transcription app?

No — transcription is only the first stage. VoiceWiz applies a configured context and (optionally) a Prompt Builder stage that reshapes the transcript into task-ready output. If you want plain transcription, Direct mode provides exactly that.

Does VoiceWiz work offline?

The local pipeline does. With a local transcription model and a local refinement model, dictations are processed entirely on your machine without a network connection. Model downloads and cloud providers (if configured) are the only steps that need the network.

Can VoiceWiz use local models?

Yes — whisper-class models and NVIDIA Parakeet for transcription, and llama.cpp or Ollama models for refinement. Cloud providers (OpenAI, Anthropic, Gemini, Groq, DeepSeek, OpenRouter) are available too, using your own API keys.

What happens to my audio and text?

Dictations and history are stored in a local database on your machine. Audio is processed by the backend you configure: local models keep everything on-device; cloud providers you configure receive the audio for transcription, as with any API-based service.

Is VoiceWiz free?

VoiceWiz will be free to access during the first month after release. Pricing after the launch period will be announced separately.

Which operating systems are supported?

Windows and Linux today. A macOS build is planned — watch the releases page.

Do I need an account?

No. There's no sign-in and no account — download the app and use it. Cloud providers, if you choose them, use your own API keys.

Can I use VoiceWiz in another language?

Whisper-class models support dozens of languages; Parakeet v3 supports 25. Set your language during the first-run wizard or in settings. Dictionary hints work per-language.