Voice AI Runtime

Start with a local voice runtime

Download Izwi, load a supported model, and make the first request on your own machine. Test the voice workflow locally, then add deployment guidance and support when you are ready for production.

First session checklist

  1. 1Install Izwi
  2. 2Pull a model
  3. 3Start the local server
  4. 4Make your first transcription request
  5. 5Explore the desktop surface

Select build

Choose the build for your machine, then use the docs to pull a model and start the local API server.

macOS

Latest release v0.1.0-beta-17 - 30.1 MB

Download for macOS

The macOS build may require manual approval in System Settings because it is currently unsigned.

Need a supported production path?

Take a validated workflow into production

Talk to Izwi when you need deployment in a customer VPC or data center, versioned releases, rollout guidance, observability, or ongoing support.

Talk to sales

Deployment service

Need a private LLM deployment?

Izwi can evaluate, deploy, and integrate an open-weight language model in your cloud account or data center. Start with the use case, target environment, and expected load.

View LLM Deployment