Claude Code takes notes for itself — and it's surprisingly good at knowing what to remember

CLAUDE.md is what you write. Auto memory is what Claude writes for itself. It's a system where Claude saves notes as it works — build commands it discovers, debugging patterns, your code style preferences — so it knows these things next session.

The files live in ~/.claude/projects//memory/ as plain markdown you can read, edit, or delete. A MEMORY.md file acts as an index, and Claude creates topic files like debugging.md or api-conventions.md for detailed notes.

The clever part: Claude doesn't save something every session. It decides what's worth remembering based on whether it would be useful in a future conversation. When you correct Claude — "no, we use pnpm here, not npm" — it's likely to remember that. The first 200 lines of MEMORY.md load into every session automatically.

You can browse and manage everything with the /memory command.