Company Brain Memory overview

A cross-run knowledge persistence layer for AI agents operating within an organization. Agent sessions produce memory snapshots — structured summaries of decisions, code patterns, architecture context, and lessons learned — that are stored in an AgentMemoryRepository. Subsequent agent sessions can query this repository to retrieve relevant organizational knowledge before starting work, reducing ramp-up time and improving consistency across independent agent runs. The memory layer supports vector similarity search, temporal filtering, and org-scoped access control.