Project Kaze — Architecture Documentation
This page is an index. The architecture is documented across 4 focused documents.
Documents
| # | Document | Scope |
|---|---|---|
| 1 | Philosophy & Vision | Core principles, AI-native paradigm, vertical-first flywheel, supervision ramp, knowledge architecture, agent safety boundaries, OpenClaw relationship |
| 2 | System Architecture | Layer model, component architecture (runtime, gateway, knowledge, agents), execution modes, data flows, API contracts, design decisions |
| 3 | Infrastructure | K8s topology, sidecars, CI/CD pipeline, GitOps, secrets management (Vault + ExternalSecrets), OpenClaw integration, networking, deployment modes |
| 4 | Non-Functional Assessment | Security posture, threat model, cost model, scalability stages, reliability, performance, compliance readiness |
Legacy Documents
The following documents contain the original design-phase thinking. They are superseded by the 4 documents above but retained for reference:
- ai-native.md — Original AI-native design concepts
- system-design.md — Original system design and layer model
- agent-model.md — Original agent model specification
- technical-design.md — Original component technical designs
- deployment.md — Original deployment architecture
- security.md — Original security design
- threat-model.md — Original threat model