Documentation

The trust boundary between AI intent and trusted physical reality.

These docs are organized around the current SeedCore slice: custody-aware digital twins, deterministic policy decisions, verification surfaces, and replayable proof.

Read them as infrastructure notes for the current trust-boundary product, not as generic autonomous orchestration material.

Documentation Paths

Primary technical entry points.

OverviewArchitecture overview

Runtime planes, actor relationships, trust boundary, and deployment model.

Open architecture page
ContractsIntent and routing references

Execution payloads, routing references, and stable decision-time control contracts.

Open router reference
MemoryOperational memory design

Memory, retrieval, provenance, and continuity notes for long-lived governed workflows.

Open memory design
Graph and RetrievalGraph-vector system notes

Supporting material for custody state, linked entities, and policy-aware retrieval surfaces.

Open graph-vector notes
Design NotesDevelopment notes and execution paths

Further documentation in the public repository.

Browse repo docs