Welcome to Hoziron
Hey there — glad you're here. Whether you're setting up your first instance, deploying to production, or building custom agent competencies, this documentation has you covered.
Hoziron is an enterprise AI agent infrastructure platform. It gives you a model-agnostic runtime for deploying, managing, and orchestrating AI agents with built-in data sovereignty, compliance guardrails, and multi-provider support.
Where should I start?
That depends on what you're trying to do:
| I want to... | Start here |
|---|---|
| Get a running instance in under 10 minutes | Getting Started |
| Deploy to production (Kubernetes, Docker, bare metal) | Deployment Guides |
| Create agents and competencies | Agent Guides |
| Connect to databases and APIs | Data Integration |
| Lock down security and auth | Security Guides |
| Understand the architecture | Architecture |
| Fix something that's broken | Troubleshooting |
| Look up a CLI command or API endpoint | Reference |
| Get quick answers | FAQ |
Documentation structure
- Getting Started — Your first 10 minutes. Install, configure, run your first agent.
- Guides — Task-oriented walkthroughs organized by topic: deployment, security, agents, competencies, workflows, data integration, providers, and catalog.
- Reference — Technical lookup: CLI commands, API endpoints, config file schema, error codes.
- Architecture — System design, data flow, object model, deployment topology.
- Internals — Deep dives into subsystems: provider routing, workflow engine, PII pipeline, security model.
- Troubleshooting — Something went wrong? Diagnostics, common fixes, recovery procedures.
- FAQ — Quick answers to the questions everyone asks.
- Compliance — Regulatory alignment documentation (in progress).
- Case Studies — Real-world deployment examples (coming soon).
- Dashboard — Web UI documentation (coming with v0.7).
A note on tone
We wrote this documentation for platform engineers who need to get things done. It's meant to be clear, practical, and respectful of your time. If something's confusing or missing, that's on us — please open an issue.
Related: