What is Izwi?
Izwi is a powerful, privacy-focused audio AI platform that runs entirely on your machine. No cloud services, no API keys, no data leaving your device. Key capabilities:- Voice Mode — Real-time voice conversations with AI
- Chat — Text-based AI conversations
- Transcription — Convert audio to text with high accuracy
- Voices — Manage built-in, cloned, and designed voices
- Text-to-Speech — Generate natural speech from text
- Voice Cloning — Clone any voice from a short audio sample
- Voice Design — Create custom voices from text descriptions
- Diarization — Identify and separate multiple speakers
Quick Links
| Section | Description |
|---|---|
| Getting Started | Install Izwi and run your first command |
| Installation | Platform-specific installation guides |
| Runtime Support Matrix | Supported OS, hardware, artifact, and API surfaces |
| API Reference | Stable, preview, first-party, operator, and realtime HTTP/WebSocket APIs |
| Features | Learn about each feature in detail |
| Models | Understand and manage AI models |
| CLI Reference | Complete command-line reference |
| Troubleshooting | Common issues and solutions |
System Requirements
| Requirement | Minimum | Recommended |
|---|---|---|
| macOS | 12.0+ (Monterey) | 14.0+ (Sonoma) |
| Linux | Ubuntu 20.04+ | Ubuntu 22.04+ |
| Windows | Windows 10 | Windows 11 |
| RAM | 8 GB | 16 GB+ |
| Storage | 10 GB free | 50 GB+ free |
| GPU | — | Apple Silicon / NVIDIA GPU (see support matrix) |
Note: Izwi is optimized for Apple Silicon Macs with Metal acceleration. NVIDIA CUDA support exists in the runtime, but artifact-level support varies by source build, Docker image, and release package. See the Runtime Support Matrix.
Getting Help
- GitHub Issues — Report bugs or request features
- Discussions — Ask questions and share ideas