Getting Started
Your first 10 minutes with Hoziron. By the end of this section, you'll have a running instance with an agent responding to messages.
What you'll need
- A machine with 2+ CPU cores and 2 GB RAM (Linux, macOS, or Windows WSL2)
- One of: a provider API key (Anthropic, OpenAI, Groq) or a local Ollama instance
- Docker (optional, for container deployment)
Choose your install method
| Method | Best for | Time to first agent |
|---|---|---|
| Binary install | Local dev, bare metal servers | ~3 minutes |
| Docker | Container environments, quick eval | ~2 minutes |
| Kubernetes | Production clusters | ~10 minutes |
What happens during setup
- Install — Get the Hoziron binaries or container image
- Initialize — Creates your config directory with
config.toml,.env, anddata/ - Configure a provider — Point Hoziron at an LLM (cloud or local)
- Start — Launch the API server (
hoziron-api) - Create an agent — Spawn from a template and start chatting
Quick reference
| Command | What it does |
|---|---|
hoziron init | Initialize config directory |
hoziron status | Show platform status |
hoziron health | Quick health check |
hoziron doctor | Run diagnostics (with --repair to auto-fix) |
hoziron chat | Interactive chat with default agent |
hoziron agent list | List all agents |
Environment variables
| Variable | Purpose | Default |
|---|---|---|
HOZIRON_HOME | Config/data directory | ~/.hoziron |
HOZIRON_LISTEN | Daemon listen address | 127.0.0.1:4200 |
HOZIRON_LOG | Log level filter | info |
HOZIRON_LOG_FORMAT | Log format (text or json) | text |
Ready? Pick your install method and let's go.
Related: