Documentation
Brain architecture
The detailed architecture is awaiting the source material for this revision.
What belongs here
The forthcoming material should explain data ownership, ingestion, retrieval, source links and the boundary between Brain and its adapters. Until it arrives, this page makes no storage-engine or retrieval-pipeline claim.
Evaluate the same system
The published architecture should name the configuration tested by the benchmark. Readers should be able to connect a result to the system that produced it.