Quickstart

Use this page to get oriented quickly before reading the full usage guide.

Read the key commands

The usage guide is anchored to these CLI help surfaces:

• mullgate --help
• mullgate setup --help
• mullgate start --help
• mullgate status --help
• mullgate doctor --help
• mullgate config --help

Know the invocation forms

Mullgate has three truthful invocation forms:

1. installed mullgate command
2. packed release asset
3. contributor or source-checkout path

This documentation uses installed mullgate ... examples by default.

Simple mental model

A normal Mullgate workflow looks like this:

1. run mullgate setup
2. inspect generated configuration and exposure details
3. start the runtime with mullgate start
4. verify behavior with mullgate status and mullgate doctor
5. access routes by hostname or direct bind IP

Recommended first run

For a low-friction Linux proof, start with a local loopback setup using two routes. That gives you:

• one machine
• multiple route entrypoints
• hostname mapping through a generated hosts block
• a safe environment for verifying the CLI and runtime lifecycle

See the full examples in Usage and Setup and Exposure.

Important platform note

Linux is the fully supported runtime environment.

macOS and Windows support the CLI, config path, and diagnostic surfaces, but Docker Desktop host networking differs from Linux, so the shipped multi-route runtime should not be treated as fully equivalent there.

For the exact support posture, see Platform Support.

Suggested next pages

1. Usage
2. Setup and Exposure
3. Architecture Overview
4. Current Runtime Model
5. Multi-exit Architecture Spec
6. Publishing