Welcome to #B4arena
#B4arena is an agentic company — an organization where AI agents do the work and humans steer the direction. We build software products and services, with agents handling execution while humans (spending ≤1 hour/day) own architecture and strategy.
This knowledge base is organized in four levels, from philosophy to hands-on runbooks.
How to Read These Docs
| Level | What You'll Learn | Best For |
|---|---|---|
| L0 — Philosophy | What an agentic company is, the three feedback loops, the Constitution | Understanding why #B4arena works the way it does |
| L1 — Mental Models | SOUL, Bead, Git strategy, Constructor ↔ Editor boundary | Building a vocabulary for daily work |
| L2 — Architecture | Ludus, Beads, Arena Core, Infrastructure, Observability | Understanding how the subsystems fit together |
| L3 — Runbooks | Dev setup, local development, deployment, troubleshooting | Getting things done |
Quick Links
- Source repos: arena · ludus · infra
- Research papers: Research section
- Framework docs: Constitution · Operating Model · Agent Design