Claude keeps separate memories in different places — and which ones follow you around depends entirely on how you use it. A standalone chat, a Project chat, Claude Code in WSL, the same Claude Code on the web, Cowork — each reads and writes a different set of stores.

This tool turns that into something you can see. Check every context you actually use, and the diagram composes itself: what gets in, where it lives, and how it comes back out. Toggle the scope lens to ask “what follows me across projects?” vs. “what stays put?”, and flip between brief and full labels when you want the official Anthropic terminology spelled out.

It’s a single self-contained page — the diagram is generated live in your browser from your selection, nothing is sent anywhere. The source is on GitHub; every store, edge, and term traces to documented Anthropic behavior.