Features

• Universal Audio Capture Record mic, system audio, or any combination via BlackHole (macOS) or equivalent loopback drivers.

• Manual Control Start/Stop buttons let you define exactly the boundaries of your prompt—perfect for long, multi-sentence queries.

• One-Shot Transcription Whisper processes the entire recording in a single call—no fragmented sentences.

• Live AI Streaming Gemini’s response appears token by token, just like in ChatGPT’s streaming interface.

• Configurable Model Swap between Gemini variants via environment (no code changes).

• Zero-Install UI Bootstrapping On first run voxai auto-installs the minimal Electron UI source—included in the PyPI package—so you only ever need pip install voxai and voxai.