Get Izwi running in under 5 minutes.

Step 1: Install Izwi

Before you pick an install path, check the Runtime Support Matrix if backend support matters for your deployment. Linux and Windows release installers are CPU-only; use Docker CUDA on NVIDIA Linux hosts or source builds for custom CUDA validation.

macOS

Download Izwi_<version>_aarch64.dmg (Apple Silicon) from GitHub Releases, then:
  1. Open the downloaded .dmg file
  2. Drag Izwi.app to your Applications folder
  3. Launch Izwi from Applications
To use the izwi command in a terminal, open Settings > System > Command line tool in the app and select Install. Izwi links izwi into /opt/homebrew/bin or /usr/local/bin and may ask for an administrator password.

Linux

Download Izwi_<version>_amd64.deb from GitHub Releases, then install it (this also installs the izwi command):
The Linux release package is CPU-only. Use Docker CUDA or a source build for NVIDIA acceleration.

Windows

Download and run Izwi_<version>_x64-setup.exe from GitHub Releases, then install the izwi command from Settings > System > Command line tool.
The Windows release package is CPU-only. CUDA on Windows is currently a source-build preview path.
See Installation for detailed platform-specific instructions.

Step 2: Start the Server

Open a terminal and run:
You should see:
Leave it running and use a second terminal for the next steps. Desktop mode (opens the native app):
Web mode (opens in your browser):

Step 3: Download a Model

Izwi needs AI models to work. Download your first model:
This downloads the default text-to-speech model. Install espeak-ng first (brew install espeak-ng on macOS, sudo apt install espeak-ng on Ubuntu/Debian). View all available models:
View your downloaded models:

Step 4: Try It Out

Generate Speech

Generate speech and play it with the system player (Izwi saves the file first):

Transcribe Audio

First, download an ASR model:
Then transcribe:

Step 5: Open the Web UI

With the server running, open your browser to:
The web interface provides:
  • Voice — Real-time voice conversations
  • Chat — Text-based AI chat
  • Transcription — Upload and transcribe audio files
  • Voices — Create, preview, manage, and reuse voice assets
  • Text to Speech — Generate speech from text
  • Models — Manage your downloaded models
  • Settings — Configure theme, updates, analytics, and desktop behavior

Next Steps