Install

Install the engine and start locally.

Deploy locally with a single command. Whether you're integrating Coding Agents or running daily regression tests, the local-first architecture provides a secure and cost-effective testing experience.

Install

curl -fsSL https://get.munk.sh | sh
munk doctor
munk serve --port 16888

Requirements

macOS or Linux environment with Node and Python available

A working device or target environment for verification

A valid config file for model setup (Gemma 4 recommended)

Next steps

Use runtime-dev during active development and integration.

Read the docs for multi-modal model setup, change-aware workflows, and generalization details.