Every AI session starts knowing your architecture
I kept re-explaining the same architectural decisions to Claude Code and Cursor, session after session , the same 'why did we choose X over Y' conversation, over and over. So I decided to try to fix it.