How we work
This is the documentation of the AI OS, the operating environment we use to work with AI agents. It describes what the system is, why it exists, and how its parts fit together.
These pages are the public edition of our internal system documentation. Each page is derived from the internal source and regenerated whenever that source changes.
Where to start
- What the AI OS is: the entry-point definition, the four components, and what it isn't.
- Why it exists: the four structural gaps in raw agents-as-tools, and how each component addresses one.
- Design philosophy: a central source of truth with derived state, applied across the four components.
- Operating philosophy: the solo-studio postures the executive encodes.
- Methodology: how the documentation and its derived views are produced and kept fresh.
Who this is for
We write these pages for two readers: ourselves, returning to part of the system months later, and a peer engineer who wants to understand the approach.